On Commuting Involution Graphs of Certain
Finite Groups
A thesis submitted to the University of Manchester
for the degree of Doctor of Philosophy
in the Faculty of Science and Engineering
2017
Ali Aubad
School of Mathematics
Contents
Abstract
5
Copyright Statement
7
Acknowledgements
9
1 Preface
10
2 Background
14
2.1
Group Extensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
2.2
Commuting Graphs
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
2.3
Commuting Involution Graphs . . . . . . . . . . . . . . . . . . . . . . .
21
2.4
Computational Group Theory . . . . . . . . . . . . . . . . . . . . . . . .
27
2.4.1
Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
27
2.4.2
Computer Implementations . . . . . . . . . . . . . . . . . . . . .
28
2.4.3
Randomised Algorithms . . . . . . . . . . . . . . . . . . . . . . .
28
2.4.4
Black-Box Groups . . . . . . . . . . . . . . . . . . . . . . . . . .
28
3 Commuting Involution Graphs of Double Covers of Sym(n)
30
3.1
Double Covers of Sym(n) . . . . . . . . . . . . . . . . . . . . . . . . . .
30
3.2
x-Graph of Sym(n) and its Double Covers . . . . . . . . . . . . . . . . .
33
3.3
Conjugacy Classes of Involutions of 2. Sn . . . . . . . . . . . . . . . . . .
34
3.4
The Disc Structure of C(2. Sn , X) . . . . . . . . . . . . . . . . . . . . . .
35
3.5
The connectivity of C(2. Sn , X) . . . . . . . . . . . . . . . . . . . . . . .
37
2
3.6
The Diameter of C(2. Sn , X) . . . . . . . . . . . . . . . . . . . . . . . . .
38
4 Commuting Involution Graphs of Double Covers of Sporadic Groups
and Their Automorphism Groups
66
4.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
66
4.2
General Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
67
4.3
The Mathieu Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
71
4.4
Leech Lattice Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
73
4.5
Monster Sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
82
5 Finite Groups Of Lie-type
88
5.1
Algebraic Group Theory . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
5.2
Subgroups of Algebraic Group . . . . . . . . . . . . . . . . . . . . . . . .
89
5.3
Groups with a BN -pair . . . . . . . . . . . . . . . . . . . . . . . . . . .
90
5.4
Classification of Simple Algebraic Groups . . . . . . . . . . . . . . . . .
92
5.5
Finite Groups of Lie-type . . . . . . . . . . . . . . . . . . . . . . . . . .
93
6 Commuting Involution Graphs of Exceptional Groups of Lie-type
96
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
6.2
Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
98
6.3
CG (t)-orbits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
6.4
Disc Structure
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
7 Investigation On Commuting Involution Graphs for the Exceptional
Groups of Lie-type 2 E6 (2) and E7 (2)
115
7.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
7.2
Basic Definitions and Results . . . . . . . . . . . . . . . . . . . . . . . . 116
7.3
The Commuting Involution Graphs of 2 E6 (2) . . . . . . . . . . . . . . . 117
7.4
The Commuting Involution Graphs of E7 (2)
. . . . . . . . . . . . . . . 121
7.4.1
Semisimple Classes of E7 (2) . . . . . . . . . . . . . . . . . . . . . 126
7.4.2
Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
7.4.3
Disks Structure and Orbit Size . . . . . . . . . . . . . . . . . . . 132
7.4.4
Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.5
Commuting Involution Graph of 2A . . . . . . . . . . . . . . . . . . . . 134
7.6
Commuting Involution Graph of 2B . . . . . . . . . . . . . . . . . . . . 136
7.7
Commuting Involution Graph of 2C . . . . . . . . . . . . . . . . . . . . 140
8 Conclusions and Future Work
A
155
8.1
Conclusion
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
8.2
Potential for Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Magma Implementations
157
B Magma Codes
160
Bibliography
162
Notations and Symbols
169
Abstract
The University of Manchester
ALI ABD AUBAD
Doctor of Philosophy
On Commuting Involution Graphs of Certain Finite Groups
June 30, 2017
Assume that G is a finite group and X is a subset of G. The commuting graph,
denoted by C(G, X), has vertex set X with vertices x, y ∈ X being connected
together on the condition of x 6= y and xy = yx. In this thesis, we study these
and other related graphs for particular types of finite groups such as the double
covers of symmetric groups, double covers of certain finite sporadic simple groups
and their autmorphism groups, and several of the exceptional groups of Lie-type.
We will pay specific attention to distinguish the discs structure and the diameters
for these graphs.
5
Declaration
No portion of the work referred to in this thesis has been
submitted in support of an application for another degree or
qualification of this or any other university or other institute
of learning.
6
Copyright Statement
i. The author of this thesis (including any appendices and/or schedules to this
thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he
has given The University of Manchester certain rights to use such Copyright,
including for administrative purposes.
ii. Copies of this thesis, either in full or in extracts and whether in hard or electronic
copy, may be made only in accordance with the Copyright, Designs and Patents
Act 1988 (as amended) and regulations issued under it or, where appropriate,
in accordance with licensing agreements which the University has from time to
time. This page must form part of any such copies made.
iii. The ownership of certain Copyright, patents, designs, trade marks and other intellectual property (the “Intellectual Property”) and any reproductions of copyright
works in the thesis, for example graphs and tables (“Reproductions”), which may
be described in this thesis, may not be owned by the author and may be owned
by third parties. Such Intellectual Property and Reproductions cannot and must
not be made available for use without the prior written permission of the owner(s)
of the relevant Intellectual Property and/or Reproductions.
7
iv. Further information on the conditions under which disclosure, publication and
commercialisation of this thesis, the Copyright and any Intellectual Property
and/or Reproductions described in it may take place is available in the University
IP Policy (see http://documents.manchester.ac.uk/DocuInfo.aspx?DocID=
487), in any relevant Thesis restriction declarations deposited in the University
Library, The University Library’s regulations (see http://www.manchester.ac.
uk/library/aboutus/regulations) and in The University’s policy on presentation of Theses
Acknowledgements
In the beginning, I would like to thank my Lord for unending blessings that have been
given to me and my family. Many thanks also go to my father Abd Altai and my
mother Fadilh Altai, sisters, brothers, my uncle Ahmed Dawood and his wife Sahira
Tawfiq, for their constant encouragement and belief in me. Also I express my sincere
thanks and appreciation to my wife Arwa Abdullah for her hard work and support and
for being generally excellent.
My sincere thanks and appreciation to my supervisor Prof. Peter Rowley. Thank
you for supporting me, giving me a valuable advice, and your patience during the period
of my PhD study. A big Thank you!
I am very thankful to The Higher Committee of Education Development in Iraq for
their financial support, and everyone at the School of Mathematics at the University
of Manchester. It has been an enjoyable, favourable and interesting place to study.
Finally, I am very grateful to my group; John Ballantyne, David Ward, Paul Bradley,
Tim Crinion, Alex Mcgaw, Peter Neuhaus, Jamie Phillips, Awatef Almotairi and Daniel
Vasey. For helping me to adjust to university life and collaborating with their brilliant
ideas. Special thanks go to David Ward for his never-ending willingness to help me
with everything related to my thesis. Also thanks to the wonderful colleagues in my
office; Jacob Cable, Ulla Karhumaki and Zoltan A.Kocsis. They made the office the
best place to work.
9
Chapter 1
Preface
It is believed that studying the action of a group on a graph is one of the most effective
ways of analyzing the structure of the group. From the 1950s era, the method of making
the automorphism group of the graph embedded inside the group has provided us with
a number of significant results. Currently, many important results have been obtained
using these methods. Suppose that G is a group and X is a subset of G; the commuting
graph, denoted by C(G, X) has vertex set X and vertices x, y ∈ X are connected on
the condition of x 6= y and xy = yx. The commuting graphs were first illustrated by
Brauer and Fowler in the seminal paper [21] and they were first for giving evidence
that for a given an isomorphism type of an involution centralizer, where there are
finitely many non-abelian groups capable of containing it, up to isomorphism. These
graphs are vital for work on the Margulis-Platanov conjecture (see [59] as the graphs
mentioned in [21] have X = {G \ {1}}). The complement of these graphs is referred
to as a non-commuting graph. They also occurred in [54] in which a problem posed by
Erdos was solved by B.H. Neumann. Furthermore, as per the assumptions of Maimani,
Abdollahi and Akbari, if M is a finite group with a trivial centre and G is a finite simple
group such that the non-commuting graphs of G and M are isomorphic then G and
M are themselves isomorphic. These sorts of assumptions have been proven in many
cases especially for those that satisfy the conjecture of J. Thompson (see [2, 30, 35]).
Even while studying the finite simple groups, specifically the non-abelian simple groups,
10
Chapter 1. Preface
11
different sorts of commuting graphs have been established. For instance, Aschbacher
and Segev [6] had also used the commuting graph in their computer-free uniqueness
proof of the Lyons simple group which has as vertices the 3-central subgroups of order 3.
The commuting involution graph is considered a very extraordinary type of commuting
graph of G in which the vertex set is a conjugacy class of involutions. At first, the
commuting involution graphs were utilized by Fischer in his work when he had to
evaluate his work on the 3-transposition groups which was largely unpublished [39, 40].
The vertices were believed to be of the commuting involution graph and were termed to
be conjugate involutions in which their product of any two vertices has order at most 3.
In addition, this graph led to the construction of sporadic simple groups of Fisher i.e.
F i22 , F i23 and F i24 ′ and their detailed structures, and proof the uniqueness are given
in [9]. Soon after Aschbacher [8] discovered a condition on a commuting involution
graphs of a finite group which ensures the presence of a strongly embedded subgroup
was ensured.
Rowley, Hart (nēe Perkins), Bates, and Bundy put their efforts into studying the
commuting involution graphs and provided the diameters and disc sizes. The entire
structure of C(G, X), where X is a conjugacy class of involutions of G, a symmetric
group, a finite Coexter group, a sporadic simple group or a projective special linear
group has been scrutinized at length by this quartet ([16, 18, 19, 20]). Perkins [57] also
studied the commuting involution graphs of Affine Coexter Groups. Moreover, a wider
study remarked on the action of the symmetric group on this kind of graph and can be
found in [17, 26]. On the contrary, Jafarzadeh and Iranmanesh [47], Mohammadian,
Raja, Radjavi and Akbari [4] have also determined the number of commuting graphs
through distinct ways. In their works the commuting graphs C(G, X) for a group G has
a vertex set of X = G \ Z(G) in which two different elements in X are linked together
by an edge if they commute. Nowadays, the commuting graphs are further studied in
context of rings (for example [3, 5]). This thesis involves certain discussion regarding
the research works of Rowley, Hart, Bates and Bundy in context with the study of
commuting involution graphs for symmetric groups [20].
12
This thesis presents a sequel of sorts to the research of my supervisor Prof. Peter
Rowley, in particular his work on the commuting involution graphs of symmetric groups
and particular finite simple groups. He provides diameters and scrutinizes the discs
structure of the commuting involution graphs for such groups. Here similar work will
be investigated, but for certain types of finite groups such as the double covers of
symmetric groups, double covers of certain finite sporadic simple groups and their
automorphism groups, and several of the exceptional groups of Lie-type. The tools
used for this purpose in this project will be theoretical and computational in nature
according to the structure of the groups.
In Chapter 2, we provide a brief summary of double covers of finite groups, which
we will be principally working with in the first four chapters. This chapter will be
fundamental in laying the foundations for what will come. A survey of the present
research on commuting involution graphs will be given. Finally, general conventions
and notation will be explained in this chapter.
In Chapter 3, we will the study the commuting involution graphs of double covers
of the symmetric group of degree n. The connectivity of this graph will be shown.
Additionally we will prove that the diameter of such graphs is at most 5.
In Chapter 4, the commuting involution graphs of double covers of certain finite sporadic simple groups and their automorphism groups will be examined. Full information
about these graphs will be determined during this chapter.
In Chapter 5, we give a brief review of the groups of Lie-type, covering the construction and different widely known properties of such groups. This will be key to the
study of the commuting involution graphs of exceptional groups of Lie-type in later
chapters.
In Chapter 6, we discuss the commuting involution graphs for several of the untwisted groups and the twisted groups. The research involves analyzing the discs structures and calculating the diameters of the graph.
In Chapter 7, we examine computationally the suborbit structure of involution conjugacy classes of the exceptional groups of Lie-type 2 E6 (2) and E7 (2) over GF (2). We
Chapter 1. Preface
13
apply this input to ascertain the diameter and investigate disc structures of commuting
involution graphs.
We close this thesis in Chapter 8 by providing a brief summary of our project and
present a futuristic vision about related research. Finally, we should mention that in
most computational chapters we will use the online Atlas of Group Representations
[67] to get a group representation and we refer to it as The Online Atlas. Moreover,
the Atlas of Finite Groups Representations [31] plays an important role in this thesis
and we refer to it as The Atlas.
Chapter 2
Background
In this chapter, we will give basic theoretical results and definitions associated to group
extensions. Moreover, we supply a literature review of the contemporary research related to commuting graphs and commuting involution graphs that will be used throughout our investigation.
2.1
Group Extensions
One of the most frequently asked questions in mathematics is “if we are given a space
X, can we extend it to another space Y larger then X?” On the whole, there are two
approaches to do that:
• By embedding the space X into a space Y which contains at least one an isomorphic copy of X. Equivalently, there is an exact sequence
0→X→Y
• The other way is by covering the space X by a space Y which contains an isomorphic copy thereof as a quotient. Equivalently, there is an exact sequence
0→Y →X
14
Chapter 2. Background
15
If X is both embedded in, and covered by, Y , then we call the exact sequence split.
Now we present some basic group theoretic results related to the double covers of
a group:
Lemma 2.1.1 (Dedekind Modular Law). [10] Suppose that A, B and C are subgroups
of a group G such that B contains A. Then AC ∩ B = A(B ∩ C).
Definition 2.1.2. [44] A complement of a normal subgroup H of a group G is a
subgroup K of G such that G = HK and H ∩ K = 1.
Definition 2.1.3. [48] A short exact sequence of a groups G, H and K is a sequence
i
π
1→
− H→
− G−
→K
where π is an epimorphism and i is a monomorphism such that image i =kerπ.
Definition 2.1.4. [48] A group G is called an extension of a group H by a group K if
G contains a normal subgroup A such that H ∼
= A and G/A = K. Similarly, there is a
short exact sequence
i
π
1→
− H→
− G−
→K
Definition 2.1.5. [44] Let G be an extension of a group H by a group K. If H contains
a complement in G then, the extension G is called split and denoted by G = H : K.
Otherwise, we call G a non-split extension of the group H by the group K and denoted
by G = H.K.
Definition 2.1.6. [44] Suppose that G and H are groups and π : H → Aut(G) is a
group homomorphism. Consider the set S = {(g, a)|g ∈ G, b ∈ H}, and define a binary
operation on S as follows (g, a) (k, b) = (gk aπ , ab), for all a, b ∈ H, and for all g, k ∈ G.
Then S is a group called a semidirect product of G by H and denoted by G ×π H.
The next lemma shows that there is an equivalence between the split extension and
semidirect product:
2.1. Group Extensions
16
Lemma 2.1.7. [10] Let G be a group, H a normal subgroup of G and let K be a
complement to H in G. Let π : K → Aut(H) be the conjugation map which is defined
as follows; for k ∈ K, π(k)(h) = hk for all h ∈ H. Define δ : H ×π K → G by
(k, h)δ = kh. Then δ is an isomorphism.
The following theorem explains the extension of a normal p-subgroup.
Lemma 2.1.8. (Gaschutz Theorem)[10] Let p be a prime, V an abelian normal p −
subgroup of a finite group G, and P ∈ Sylp(G). Then G splits over V if and only if P
splits over V .
Definition 2.1.9. [71] An extension G of a group H by a group K is said to be a central
extension if there is a subgroup L of the centre of G such that L ∼
= H. Furthermore,
the central extension G of a group H by a group K is called a stem extension if L 6 G′
(where G′ is the derived subgroup of G). Moreover, the stem extension of G is called
double, triple, etc., cover if the order of H is 2, 3, etc.
Definition 2.1.10. [71] A Schur cover of a finite group G is a group homomorphism
π : K → G such that π satisfies the following conditions:
1. kerπ ≤ Z(K) ∩ K ′ , where K ′ is the derived subgroup of K.
2. K has largest size between all homomorphism satisfying (1).
The kernel of π is called the Schur multiplier. The group K is often called the Schur
cover and π is surjective.
Definition 2.1.11. [63] Let G, H be groups we call G and H isoclinic if the following
three conditions holds:
1. G/Z(G) ∼
= H/Z(H).
2. The derived subgroup of G is isomorphic to the derived subgroup of H.
3. if a1 Z(G), a2 Z(G) ∈ G/Z(G) with image b1 Z(H), b2 Z(H) under the isomorphism
−1
−1 −1
(1) then a1 a2 a−1
1 a2 has image b1 b2 b1 b2 under the isomorphism (2).
Chapter 2. Background
17
The relation between the covers and the isoclinic of a finite groups can be seen in
the following result:
Theorem 2.1.12. [63] All the covering groups for a given finite group are mutually
isoclinic.
In order to illustrate the above results and definitions we provide the reader the
following examples:
Examples 2.1.13.
• Let G ∼
= H × K, is a direct product of H and K then G is an extension of H by
K or K by H. This extension is called the trivial extension.
• Suppose that V is a vector space over the field L. Then the general linear group
GL(V ) is a central extension of L∗ = L\{0} (the multiplicative group of L) by
projective linear group P GL(V ).
• The symmetric group Sn and the dihedral group Dn are extensions of An and Zn
by Z2 , respectively.
• The quaternion group Q8 , Z8 and Z4 × Z2 are extensions of Z2 by Z4 , since there
are no subgroups H, K ∈ Z8 or quaternion group Q8 such that H ∼
= Z4
= Z2 , K ∼
with trivial intersection. Therefore, the extensions are non-split in this case, while
the extension is split in Z4 × Z2 .
2.2
Commuting Graphs
In this section, we will give reviews on the commuting graph C(G, X), including definitions and some essential results related to these kind of graphs. First we will give a
summary about graph theory to consolidate our subsequent work follow by the definition of the commuting graph.
Let Γ(Ω, E) be a graph with edge set E and vertex set Ω, then Γ(Ω, E) is called
an undirected graph if whenever (x, y) ∈ E then (y, x) ∈ E for all x, y ∈ Ω, and with
2.2. Commuting Graphs
18
loops if there is x ∈ Ω such that (x, x) ∈ E. Suppose that d is a distance between any
two elements x, y ∈ Ω. We define d(x, y) = j where j is the length of the shortest path
between x, y. Hence d is the standard distance metric on Γ(Ω, E). The ith disc of the
element x ∈ Ω is defined by
∆i (x) = {y ∈ Ω|d(x, y) = i}.
If the length of the 1st disc has the same value for all x ∈ Ω, then the graph Γ(Ω, E)
is called a regular graph, and the quantity |∆1 (x)| is call the valency of a regular graph
Γ(Ω, E). The graph Γ(Ω, E) is called connected if there is a path between any two
vertices in Ω. Finally, if Γ(Ω, E) is a connected graph, the diameter of Γ(Ω, E) is
denoted by
DiamΓ(Ω, E) = maxx∈Ω {i|∆i (x) 6= φ and ∆i+1 (x) = φ}.
Definition 2.2.1. [21] Suppose that G is a group and let X be a nonempty subset of
G. The commuting graph of G, denoted by C(G, X), has X as its vertex set with two
distinct elements of X joined by an edge whenever xy = yx.
The next two theorems, proved by A.Iranmanesh and A.Jafarzadehin in [47], where
X taken to be Sn \{1}(n > 2) in 1. and X = An \{1}(n ≥ 5) in 2:
Theorem 2.2.2. [47]
1. The commuting graph C(Sn , X) is connected if and only if both n and n − 1 are
not prime numbers and in this case Diam(C(Sn , X) 6 5) and the bound is sharp.
2. The commuting graph C(An , X) is connected if and only if neither of n, n − 1 and
n − 2 is a prime number and in this case Diam(C(An , X) < 6) and the bound is
sharp.
The following theorem formulated some results concerning the commuting graph of
the dihedral group with a variety of different subsets X:
Chapter 2. Background
19
Theorem 2.2.3. [29] Let G = C(D2n , X), where X is a subset of D2n and n > 3. Then
the following are true:
1. If X is an abelian subgroup of D2n , then Diam(G) = 1.
2. If X is a non-abelian subgroup of D2n , then Diam(G) = 2.
3. If X = D2n \Z(D2n ), then Diam(G) = ∞ ( the graph is disconnected).
Let X be a G-conjugacy class of elements of order 3 in Sn , thus X = tG , where
t = (1, 2, 3)(4, 5, 6)(7, 8, 9) . . . (3r − 2, 3r − 1, 3r) of cycle type 1n−3r 3r . Then Athirah
Nawawi and Peter Rowley proved the following results:
Theorem 2.2.4. [53]
1. If n > 8r, then Diam(C(Sn , X)) = 2.
2. If 6r < n < 8r, then Diam(C(Sn , X)) = 3.
3. If r > 1 and n = 6r, then Diam(C(Sn , X)) 6 4.
Now we indicate important results on the commuting graph for finite soluble groups.
Theorem 2.2.5. [56] Suppose that G is a finite soluble group with trivial centre and
X = G\Z(G). Then
1. C(G, X) is disconnected if and only if G is a Frobenius group or a 2-Frobenius
group.
2. If C(G, X) is connected, then C(G, X) has diameter at most 8.
Furthermore, there exist soluble groups G with trivial centre such that C(G, X)
has diameter 8.
Throughout the next results, valuable information about the diameter of the commuting graph C(G, X) is given, where G is a finite non-commutative semigroup and
X = G\Z(G):
2.2. Commuting Graphs
20
Theorem 2.2.6. [11] For every n > 2, there is a semigroup G such that the diameter
of C(G, X) is n.
The complement of the commuting graph is called a non-commuting graph if X =
G\Z(G) and it is denoted by Γ(G). Paul Erdos in 1975 [54] was the first author who
studied this type of the graph. After that much research appeared about the noncommuting graph (for example [1, 2, 35, 62, 69]).
Now we will review some of the results related to non-commuting graph Γ(G), for
a variety of groups.
Proposition 2.2.7. [52] For every group G, the non-commuting graph Γ(G) is connected.
Theorem 2.2.8. [52] Let G be one of the following groups:
1. G = Sn , symmetric group of degree n where n > 3;
2. G = An , alternating group of degree n where n > 4;
3. G is a simple group of Lie type with(t(G) > 2 is number of connected components
of Γ(G)).
If H is a group such that Γ(G) ∼
= Γ(H). Then |H| = |G|.
Proposition 2.2.9. [34]
1. Let G be a non-abelian group, Γ(G) be the non-commuting graph of G and let
g ∈ G\Z(G) be an element of order pqr where p, q and r are distinct primes with
p < q < r and q ∤ r − 1 and p ∤ r − 1. If H is a group such that Γ(G) ∼
= Γ(H),
then |G| = |H|.
2. Let G be a non-abelian group, Γ(G) be the non-commuting graph of G and let
g ∈ G\Z(G) be an element of order pn q where p, q are primes where q < p . If H
is a group such that Γ(G) ∼
= Γ(H), then |G| = |H|.
Theorem 2.2.10. [62] Let G and H be two finite non-abelian nilpotent groups with
irregular non-commuting graphs such that Γ(G) ∼
= Γ(H). Then |G| = |H|.
Chapter 2. Background
21
The next couple of examples are established to clarify the previous results:
Examples 2.2.11.
1. Let G = D16 , the dihedral group of order 16, and X = G \ Z(G). Then C(G, X) is
a disconnected commuting graph. Indeed, C(G, X) has 5 connected components
four of them have length 2 and one has length 6.
2. Let G = A12 , the alternating group of degree 12, and let X = tG , where
t = (12)(34). Then C(G, X) is a connected commuting involution graph, with
DiamC(G, X) = 2. Also one can find by using a computational approach that the
disc structure of t is as follows
35
268
1
t
1
∆1 (t)
36
232
∆2 (t)
Where |∆1 (t)| = 268 and |∆2 (t)| = 1216. Also for x ∈ ∆1 (t) and y ∈ ∆2 (t)
the number of elements in ∆1 (t) commute with t, x, y respectively equal 268, 35,
232 and the number of elements in ∆2 (t) which commute with x, y is 232, 36
respectively.
2.3
Commuting Involution Graphs
In this section we will provide the mathematical definition of the commuting involution
graph. Moreover, we furnish an overview about the latest results related to these
graphs.
A very interesting special case of the commuting graph occurs when the subset X
of G is a conjugacy class of involutions. In this case the graph C(G, X) is called the
commuting involution graph and this type of graph has a vital role in determining the
order of the product of two elements in X. Therefore it is clear that in this graph
∆1 (x) = {y|y ∈ X ∩ CG (x) and x 6= y}. Due to the transitive action of G on X by
conjugation, it is clear that C(G, X) is a regular, undirected graph without loops.
2.3. Commuting Involution Graphs
22
This graph was first considered by Peter Rowley and his PhD students and postdoctoral researchers (Chris Bates, David Bundy and Sarah Perkins) in their paper
Commuting involution graphs for symmetric groups [19]. In that paper they proved
the following results:
Theorem 2.3.1. [19] Let a = (1 2)(3 4) . . . (2m − 1 2m) be an involution in Sn and let
X = aG . Then C(Sn , X) is disconnected if and only if n = 2m + 1 or n = 4 and m = 1.
Theorem 2.3.2. [19] Let a = (1 2)(3 4) . . . (2m − 1 2m) be an involution in Sn and let
X = aG . Suppose that C(Sn , X) is connected. Then one of the following holds:
i If 2m + 2 = n ∈ {6, 8, 10} then DiamC(Sn , X) = 4, otherwise
ii DiamC(Sn , X) 6 3.
Corollary 2.3.3. [19] Let a = (1 2)(3 4) . . . (2m−1 2m) be an involution in H ∼
= Alt(n),
the alternating group of degree n, and let X = aH . If C(H, X) is connected, then either
DiamC(H, X) 6 3 or 2m + 2 = n ∈ {6, 10} and DiamC(H, X) = 4.
Definition 2.3.4. [44] A group W is said to be a Coxeter group if the following
conditions holds:
• W generated by distinct involution wi , 1 6 i 6 m
• If wi wi has order kij , then the relations
(wi wj )kij = 1 1 6 i, j 6 m.
are a complete set of defining relations for W .
The next result provides detail about diameter and connectivity of the commuting
involution graph C(G, X) where G is a finite Coxeter group.
Theorem 2.3.5. [18] Let G be a finite Coxeter group and X a conjugacy class of
involutions in G.
Chapter 2. Background
23
1. If G is of type Bn or Dn , then C(G, X) is either disconnected or connected of
diameter at most 5, with equality in exactly one case.
2. If G is of type E6 , then C(G, X) is connected of diameter at most 5.
3. If G is of type E7 or E8 , then C(G, X) is connected of diameter at most 4.
4. If G is of type F4 , H3 or H4 , then either C(G, X) is disconnected or connected of
diameter 2.
5. If G is of type In , then C(G, X) is disconnected.
Proof.
This is an abridged version of the theorem. See [18] for more information about the
proof of this theorem.
Soon thereafter, Perkins [57] considered the case of a class of infinite groups and
their commuting involution graphs. She proved the following results:
Theorem 2.3.6. (Perkins) [57] Let G be an affine Coxeter group of type Ân , and X
a conjugacy class of involutions of G. Then C(G, X) is disconnected or is connected of
diameter at most 6.
Proof. For full details about the steps of the proof we refer to [57].
The next results provide a brief review for the diameter and disc structure of commuting involution graphs of the special linear and projective special linear groups over
several fields as mentioned in [20], and then for a 3-dimensional projective special unitary group as given in [36]. The full proofs and details about these results can be found
in [36, 37, 38] and [20] respectively,
Theorem 2.3.7. (A. Everett)[36, 37] Suppose G ∼
= SU 3(q), the 3-dimensional projective special unitary group over the finite field of q elements such that q = ps , p is
an odd prime and s ∈ N, and X a G-conjugacy class of involutions. Let t ∈ X. Then
C(G, X) is connected of diameter 3, with disc sizes
2.3. Commuting Involution Graphs
24
|∆1 (t)| = q(q − 1);
|∆2 (t)| = q(q − 2)(q 2 − 1); and
|∆3 (t)| = (q + 1)(q 2 − 1).
Theorem 2.3.8. (A. Everett, P.Rowley )[38] Suppose H ∼
= Sp(4, q), q = pa and p
a prime. Let V be the natural (symplectic) GF(q) H-module, and set G ∼
= H/Z(H).
So G ∼
= P Sp(4, q) the finite 4-dimensional projective symplectic group. Set V (x) =
{v ∈ V |(v, v x ) = 0}. Then we have
• If p = 2 and X = {x ∈ G|x2 = 1, dimCV (x) = 3} or X = {x ∈ G|x2 =
1, dimCV (x) = 2, V (x) = V } with t ∈ X. Then C(G, X) is connected of
diameter 2, with disc sizes being
|∆1 (t)| = q 3 − 2; and
|∆2 (t)| = q 3 (q − 1).
• If p = 2 and X = {x ∈ G|x2 = 1, dimCV (x) = 2, dimV (x) = 3} with t ∈ X.
Then C(G, X) is connected of diameter 4, with disc sizes being
|∆1 (t)| = q 2 (2q − 3);
|∆2 (t)| = 2q 2 (q − 1)2 ;
|∆3 (t)| = 2q 3 (q − 1)2 ; and
|∆4 (t)| = q 4 (q − 1)2 .
• If p is odd and X is a G-conjugacy class of involution whose elements are the
images of an involution in H. Let t ∈ X. Then C(G, X) is connected of diameter
2, with disc sizes being
|∆1 (t)| = 21 q(q 2 − 1); and
|∆2 (t)| = 12 (q 4 − q 3 + q 2 + q − 2).
Chapter 2. Background
25
• If p is odd and X is a G-conjugacy class of involution whose elements are the
images of an elements of H of order 4 which square to the non-trivial element of
Z(H). Let t ∈ X.
– If q ≡ 3 (mod 4). Then C(G, X) is connected of diameter 3. Furthermore,
|∆1 (t)| = 12 q(q 2 + 2q − 1);
|∆2 (t)| =
1
16 (q
+ 1)(3q 5 − 2q 4 + 8q 3 − 30q 2 + 13q − 8); and
|∆3 (t)| =
1
16 (q
− 1)(5q 5 − 4q 4 − 2q 3 + 4q 2 + 5q + 5).
– If q ≡ 1 (mod 4). Then C(G, X) is connected of diameter 3. Furthermore,
|∆1 (t)| = 12 q(q 2 + 1);
|∆2 (t)| =
1
16 (q
− 1)(3q 5 − 6q 4 + 32q 3 − 10q 2 − 27q − 8); and
|∆3 (t)| =
1
16 (q
− 1)(5q 5 + 22q 4 − 8q 3 + 34q 2 + 51q + 24).
Theorem 2.3.9. (Bates, Bundy, Perkins, Rowley) [20] Suppose G ∼
= L2 (q), the
2-dimensional projective special linear group over the finite field of q elements, and X
a G-conjugacy class of involutions with t ∈ X.
(i) If q is even, then C(G, X) consists of q + 1 cliques each with q − 1 vertices.
(ii) If q ≡ 3 (mod 4), with q > 3, then C(G, X) is connected and Diam C(G, X)= 3.
Furthermore,
|∆1 (t)| = 12 (q + 1);
|∆2 (t)| = 41 (q + 1)(q − 3); and
|∆3 (t)| = 41 (q + 1)(q − 3).
(iii) If q ≡ 1 (mod 4), with q > 13, then C(G, X) is connected and Diam C(G, X) = 3.
Furthermore,
|∆1 (t)| = 12 (q + 1);
2.3. Commuting Involution Graphs
26
|∆2 (t)| = 14 (q − 1)(q − 5); and
|∆3 (t)| = 41 (q − 1)(q + 7).
Theorem 2.3.10. (Bates, Bundy, Perkins, Rowley) [20] Suppose G ∼
= SL3 (q),
and X is the G-conjugacy class of involutions. Let t ∈ X. Then C(G, X) is connected
with DiamC(G, X) = 3 and the following hold.
(i) If q is even, then
|∆1 (t)| = 2q 2 − q − 2;
|∆2 (t)| = 2q 2 (q − 1); and
|∆3 (t)| = q 3 (q − 1).
(ii) If q is odd, then
|∆1 (t)| = q(q + 1);
|∆2 (t)| = (q 2 − 1)(q 2 + 2); and
|∆3 (t)| = (q + 1)(q − 1)2 .
Theorem 2.3.11. (Bates, Bundy, Perkins, Rowley) [20] Let K be a (possibly
infinite) field of characteristic 2, and suppose that G ∼
= SLn (K) and X a G-conjugacy
class of involutions containing t. Also let V denote the natural n-dimensional KGmodule, and set k = dimK [V, t], where [V, t] = {v ∈ V |v t = v}.
(i) If n > 4k, then DiamC(G, X) = 2.
(ii) If 3k 6 n < 4k, then DiamC(G, X) 6 3.
(iii) If 2k < n < 3k, or k is even and n = 2k, then Diam C(G, X) 6 5.
(iv) If n = 2k and k is odd, then Diam C(G, X) 6 6.
Chapter 2. Background
27
The structure of the commuting involution graph of the sporadic simple groups
has taken a large space in recent studies. The following results are related to these
kinds of graphs. For a deeper understanding to the notation of the conjugacy classes
of involutions in these results we refer the reader to the Atlas[31].
Theorem 2.3.12. (Bates, Bundy, Hart, Rowley; Rowley, Taylor; Rowley)[16]
[60] [64] Let K be a sporadic simple group and K 6 G 6 Aut(K). Let X be a conjugacy
class of involutions in G.
(i) For (K, X) not equal to (J4, 2B),(F i′24 , 2B),(F i′24 , 2D),(B, 2C),(B, 2D) or (M, 2B),
the diameter of C(G, X) is at most 4, with equality in precisely four cases.
(ii) For (K, X) equal to (J4, 2B), (F i′24 , 2B), (F i′24 , 2D), the diameter of C(G, X) is
3.
(iii) For (K, X) equal to (M, 2B) the diameter of C(G, X) is 3.
Proof. Bates, Bundy, Hart and Rowley [16] provided the proof of part (i). Rowley
and Taylor [64] gave the proof of part (ii). Finally, an unpublished manuscript of
Rowley [60] established part(iii).
2.4
Computational Group Theory
This section will clarify the general concepts concerning the computational approach
that will be utilized to study commuting involution graphs and relevant group theory.
2.4.1
Algorithms
The “pseudocode” format used in [45] will be applied to display many of our algorithms. Therefore, we consider standard coding constructs for example, statements like
“if ...then” or loops such as “for”. However, the statements will be established in
ordinary mathematical language instead of complicated programming language. Also,
for a variable y taking a value x we use the notation x 7−→ y. The application of the
algorithms can be found in the fourth chapter onwards.
2.4. Computational Group Theory
2.4.2
28
Computer Implementations
In computational group theory, the computer algebra systems Magma [23] and GAP
[41] are considered to be a most commonly utilized. In most steps of our algorithms,
we mainly use Magma, whilst GAP is used for the implementation associated with the
character table of the group. This is because GAP provides a comprehensive library of
character tables produced from Atlas [31] and is computationally efficient. Moreover,
the electronic files associated to this thesis are all set up in Magma format. Appendices
A and B have information about Magma functions.
2.4.3
Randomised Algorithms
The search for elements possessing certain properties may be impossible in some cases,
especially if the search process occurs within groups having a large size. Thus the best
available solutions will be found by finding random elements inside such groups and
checking if they have the qualities required. This method is known as a “Randomised
Algorithm” and the main role for this kind of algorithm will be looking for the orbits
that form the foundation on which to build the commuting involution graphs. Implementation of the procedures associated with randomised algorithms will be via Magma
packages.
2.4.4
Black-Box Groups
The concept of the black-box group was first introduced in 1984 by Babai and Szemerdi
see[12], with the purpose of checking basic group theory properties such as commutativity or finding random elements of the group. The elements of the black-box group
are designated (not always uniquely) by bit-strings of uniform length N , also there
exist black-boxes (oracle) used to perform the group operations. Let g, h be strings in
G. Then the black-boxes calculate gh, g −1 and checks whether h = g, each operative
performed during a specified period of time. As an example for a black-box group we
explain the Bray method established in [22] which aim to calculate the centralizer of
an involution. This method is based on the following result proved in [22].
Chapter 2. Background
29
Lemma 2.4.1. [22] Let G be a group and t ∈ G be an involution. For any g ∈ G. Let
−1
m be the order of ttg . Then if m is even, (ttg )m/2 , (ttg )m/2 ∈ CG (t) whereas if m is
odd, g(ttg )m−1/2 ∈ CG (t).
Proof. See [22, section 2.2].
According to the above lemma we build Algorithm 1.
Algorithm 1
.
Input: G a black-box group, t an involution in G;
i: g −→ Random(G), set n =Order(ttg ).
ii: if n is even then
−1
c1 = (ttg )n/2 and c2 = (ttg )m/2 else;
iii: c1 = g(ttg )m−1/2 and c2 = t.
Output: c1 , c2 ∈ CG (t).
The implementation of Algorithm 1 involves choosing a random element g ∈ G,
computing inverses and products when constructing the element ttg and the centralizing
elements, which involves computing the order of ttg . Thus Algorithm 1 is black-box.
Chapter 3
Commuting Involution Graphs of
Double Covers of Sym(n)
In this chapter we investigate the commuting involution graphs of double covers of symmetric groups by studying the structure of the graph. We also analyze the connectivity
and diameter of the graph. I. Schur [61], proved that for n > 4, the symmetric group
Sn has two proper double covers namely Sn+ and Sn− and they are isomorphic only when
n = 6. Furthermore, the Schur multiplier of these covers is a cycle group of order 2. We
will set t = −1 to be the generator of the centre for this 2-group ( Schur multiplier) so
that t is an involution in 2. Sn and 2. Sn / < t >∼
= Sn . In order to cope with C(2. Sn , X)
where X is an involution class in 2. Sn we need to have enough information about the
product of involutions of 2. Sn . The next section will deal with this for both covers of
Sn .
3.1
Double Covers of Sym(n)
Definition 3.1.1. (Sn− )[24] It is the double cover of Sn which lifts a transposition of
Sn to an element of order 4 in Sn− . This cover of Sn has generators t, r1 , r2 , . . . , rn−1
and relations:
t2 = 1;
30
Chapter 3. Commuting Involution Graph of 2. Sn
31
tri = ri t, ri2 = t for i = 1, 2, . . . , n − 1;
(rj rj+1 )3 = t for j = 1, 2 . . . , n − 2; and
rk rh = t rh rk
for |h − k| > 1 and h, k = 1, 2, . . . , n − 1.
Definition 3.1.2. (Sn+ )[24] It is the double cover of Sn which lifts a transposition of
Sn to an element of order 2 in Sn+ . This cover of Sn has generators t, r1 , r2 , . . . , rn−1
and relations:
t2 = 1;
tri = ri t, ri2 = 1 for i = 1, 2, . . . , n − 1;
(rj rj+1 )3 = 1 for j = 1, 2, . . . , n − 2; and
rk rh = t rh rk for |h − k| > 1 and h, k = 1, 2, . . . , n − 1.
To deal with a calculations of the Schur double covers we will use the method of
Conway and others at Cambridge. This first appears in the Atlas [31]. This method is
summarized in papers by David B. Wales[68] and J. Brinkman [24]. In this the elements
of 2. Sn are products of the form ±[σi ], where σi are disjoint cycles in Sn . These elements
are the two lifts of [σi ] in 2. Sn . Consider the complex algebra of the Clifford Algebra
C(Ω) as mentioned in [24] and [68] (a Clifford Algebra is a unital associative algebra
generated by a vector space with a quadratic form) where Ω = {1, 2, · · · , n} ∪ δ. Which
has the following generators A1 , A2 . . . , Aδ such that A2i = 1 in Sn+ ( A2i = −1 in Sn− )
and Ai Aj = −Aj Ai for i 6= j. The subgroups of these complex algebra generated by
√
√
√
(A1 − A2 )/ 2, (A2 − A3 )/ 2, . . . , (An − Aδ )/ 2
√
is isomorphic to 2. Sn . By denoting ai with (Ai − Aδ )/ 2, such that i′ s are distinct in
Ω \ {δ}. We obtain immediately the following relations
(ai )2 = 1 in Sn+ ((ai )2 = −1 in Sn− ) and a1 a2 · · · am a1 = (−1)m+1 a2 a3 · · · am a1 a2 .
For more information regarding Clifford Algebra and their representation we refer the
interested reader to [42] and [71].
The idea of this method relies mainly on the following definition:
Definition 3.1.3. [68] [24] For the elements a1 , a2 , . . . , am we define
3.1. Double Covers of Sym(n)
32
[a1 a2 . . . am ] = a1 a2 . . . am a1
We call ±[a1 a2 · · · ak ] signed cycles in 2. Sn . Each is a lift of the cycle (a1 a2 · · · ak ) in
Sn . Similarly for any permutation k in Sn , we have the signed permutations ±kn in
2. Sn . In fact each ai corresponds to an element of a subgroup of a Clifford algebra
which is isomorphic to 2. Sn .
The following rules are sufficient to enable the calculation of products of disjoint
signed cycles in 2. Sn (these appear as 2.3 and 2.4 in [68] and 3.1 in [24]).
1. [ai ] = t in Sn− and [ai ] = 1 in Sn+ .
2. [a1 a2 · · · am ] = (t)m+1 [a2 a3 · · · am a1 ].
3. [a1 a2 · · · am−1 ]am = (t)m am [a1 a2 · · · am−1 ].
To be familiar with the above relations we form the following basic example in 2. Sn− :
Example 3.1.4. [12][12] = 121121 = 12[1]21 = t1221 = t1[2]1 = t2 [1] = t.
The next lemma will give way for explicit computation inside 2. Sn .
Lemma 3.1.5. [68] The following relations hold in 2. Sn .
1. [i j][k l] = −[i j] where i, j, k, l are distinct.
2. [i j][j k] = −[i k] where i, j, k are distinct.
3. [i j][i j] = −[j i] = [i j] where i, j are distinct.
4. [i j]k = (sgn k)[u v] where k is a permutation in Sn which maps i to u and j to
v and sgn k is the sign of k considered as a permutation in Sn .
5. [i j][i k] = [i j k] = (−[i k j])2 where i, j, k are distinct.
Proof. The relations hold in case of the cover Sn− see [68] and for the cover Sn+
may be proved in a similar way.
Chapter 3. Commuting Involution Graph of 2. Sn
33
For a transposition (ri rj ) ∈ Sn , we call ±[ri rj ] a signed transposition in 2. Sn . The
following proposition shows when the product of disjoint signed transposition for both
covers of 2. Sn has order 2.
Proposition 3.1.6. [25]. let X be the product of m disjoint signed transposition in
2. Sn . Then
1. if 2. Sn ∼
= Sn− and is m congruent to 0 or 3 mod 4, then X has order 2, otherwise
X has order 4.
2. if 2. Sn ∼
= Sn+ and is m congruent to 0 or 1 mod 4, then X has order 2, otherwise
X has order 4.
3.2
x-Graph of Sym(n) and its Double Covers
.
Suppose that G = 2. Sn and a is an involution in G. Set X = a(2 Sn ) , the conjugacy
class of involution for a. We may assume without loss of generality that
a = [1 2][3 4] · · · [2m − 1 2m][2m + 1][2m + 2] . . . [n],
where m is congruent to 0 or 3 mod 4 in Sn− and congruent to 0 or 1 mod 4 in Sn+ (see
Proposition 3.1.6). For x ∈ G we define f ix(x) to be the set of fixed point of x on the
set Ω = {1, 2, . . . n} . The support of x, which denoted by supp(x) is defined to be the
set Ω \ f ix(x). One can see that |f ix(a)| = n − 2m, thus |supp(a)| = 2m.
Definition 3.2.1. (x-Graph in Sn ) [19]
Let Sn be the symmetric group of degree n, acting by usual manner on the set
Ω = {1, 2, · · · , n}. Let b = (1 2)(3 4) · · · (2m − 1 2m)(2m + 1)(2m + 2) . . . (n) be an
involution in Sn and Y = bSn . Assume that V = {{1, 2}, {3, 4}, · · · , {2m−1, 2m}, {2m+
1}, {2m + 2}, . . . , {n}} hence V is the set of hbi-orbit of Ω, and suppose that x ∈ Y . We
define the x-graph, denoted by Gx , to be the graph which has V as its vertex set and
two subset v1 , v2 in V are connected by an edge if there is α1 ∈ v1 and α2 ∈ v2 with
α1 6= α2 , and α1 x = α2 . We call the vertices generated by a transposition in Sn black
3.3. Conjugacy Classes of Involutions of 2. Sn
34
vertices and denote then by (•). Similarly we call the vertices generated by element
which lie in f ix(b) white vertices and denote then by (◦). An example for an x-graph
let n = 8, m = 3 , b = (12)(34)(56) and x = (13)(24)(76). Then Gx is
.
In general we denote the x-graph Gx by Gxb , that is the graph that has the orbits of
Ω under hbi as a vertex set and the orbits of Ω under hxi as defining its edges. Now
′
let x ∈ X. We may assume that Gxa = Gxa′ , where x′ , a′ are the image of x, a by
the covering map respectively. Therefore, it is obvious that x and tx have the same xgraph in 2. Sn . An example for this case, let n = 10, m = 4, a = [12][34][56][78][9][10]and
x = [14][23][56][910], then the x-graph has the following connected components
. We should note that if Gx has a connected component
of the form
. Then Gx will produce a connected component of the form
corresponding to the component
,
.
The next lemma determines the possible connected components of Gx .
Lemma 3.2.2. [19] Let x ∈ X. Then the x-graph has one of the following connected
components:
1-
,
2-
,
3-
3.3
,
,
,
,
(all black vertices);
(black and white vertices);
(all white vertices).
Conjugacy Classes of Involutions of 2. Sn
The next theorem is essential to determine the size of conjugacy classes in 2. Sn .
Theorem 3.3.1. [46] Let H be a group with a central subgroup Z = hzi of order 2 and
θ : H −→ H/Z be the natural homomorphism. For any conjugacy class C̄ of H̄ = H/Z,
the inverse image θ−1 (C̄) is either a conjugacy class in H or a union of two classes in
H. The latter case arises precisely when there is no element of θ−1 (C̄) conjugate to z
times itself.
Chapter 3. Commuting Involution Graph of 2. Sn
35
The following proposition pins down the size of conjugacy classes of involution in
2. Sn .
.
Proposition 3.3.2. For n > 4. Let a = [1 2][3 4][5 6] · · · [2m − 1 2m] and X = a2 Sn a
conjugacy class of involution in 2. Sn . Let ā be the image of a under the covering map
in Sn . If s is the size of the class āSn then 2s is the size of X.
Proof.
By Proposition 3.1.6 we have m congruent to either 0, 3 or 1 modulo 4 and
n = 2m + r where r = |f ix{a}|. Using the above theorem it is enough to prove
that ta and a are conjugate.
Therefore, if m is congruent to 0 mod 4 then set
c = [12] so ac = [12][12][34] . . . [2m − 1 2m]([12])−1 , but m is even we get ac =
t[12][12]([12])−1 [34] . . . [2m − 1 2m], thus ac = ta. If m = 1 set c = [3 4] then part
1 of Lemma 3.1.5 shows that ac = ta. The remaining cases are m congruent to 1 or 3
mod 4 and m 6= 2, in this case put c = [13][24], then ac = [1 3][2 4][12][34] . . . [2m −
1 2m]([1 3][2 4])−1 , this lead to ac = [1 3][2 4][12][34]([2 4])−1 [1 3]−1 . . . [2m − 1 2m], so
that ac = t[1 3][2 4]([2 4])−1 [1 3]−1 [12][34] . . . [2m − 1 2m]. Therefore, ac = ta .
3.4
The Disc Structure of C(2. Sn , X)
The structure of the x-graph of the element x ∈ X plays a crucial role in determining
whether x commutes with a or not. We know that, if x ∈ ∆1 (a) ∪ {a}, then this
implies x̄ ∈ ∆1 (ā) ∪ {ā} in Sn , where x̄, ā are the image of x, a under the covering map
respectively. Therefore, information about the discs structure for C(Sn , X̄) will be of
great importance when studying discs structure for C(2.Sn , X), where X̄ is the image
of the class X under the cover map. The next lemma provides us enough detail about
the discs structure for C(Sn , X̄).
Lemma 3.4.1. [19] Let x̄ ∈ X̄. Then x̄ ∈ ∆1 (ā) ∪ {ā} if and only if each connected
component of Gx̄ is one of
,
,
, or
.
3.4. The Disc Structure of C(2. Sn , X)
36
For x ∈ X the next lemma shows what are the necessary and sufficient conditions
on Gx to know that x commutes with a.
Lemma 3.4.2. Let x ∈ X. Then x ∈ ∆1 (a) ∪ {a} if and only if Gx has the following
connected components:
(i)
even number,
even number,
even number +
;
(ii)
odd number,
odd number,
odd number +
;
(iii)
even number +
(iv)
even number,
(v)
.
;
even number +
; or
Proof. By Lemma 3.4.1 we only need to show that the necessary part of the
argument. In order to prove (iii) we may assume that k is the number of cycles of length
2 in Gx (
) . So without loss of generality , we may let a = [1 2][3 4] · · · [2m−1 2m]
and x = [1 3][2 4] · · · [4k −3 4k −1][4k −2 4k][4k +1 4k +2][4k +3 4k +4] · · · [2m−1 2m].
Since [[a1 a2 ][a3 a4 ], [a1 ai ][aj ak ]] = t, we have [a, x] = tk where {i, j, k} = {2, 3, 4} and
a1 , a2 , a3 , a4 ∈ supp(x). This proves (iii). To prove (iv) we may assume that l is the
number of two connected white vertices in Gx ( ). So without loss of generality we let
x = [2m + 1 2m + 2][2m + 3 2m + 4] · · · [2(l + m) − 1 2(l + m)][2l + 1 2l + 2][2l + 3 2l +
4] · · · [2m − 1 2m][1][2] . . . [2l][2(l + m) + 1] . . . [n]. Also since [[a1 a2 ], [a3 a4 ]] = t for
different ai s. Therefore, [a, x] = tl and this proves (iv). Furthermore, (i),(ii) follow for
the proof of (iv) and (iii). Also (v) occurs when a = x.
The next example show how Lemma 3.4.2 works
Example 3.4.3. Let n = 12, and m = 4, then a = [12][34][56][78]. Suppose that
x = [13][24][67][58], y = [12][34][67][910] and w = [13][24][910][1112]. Then Gx , Gy ,
and Gw , have the following connected components, respectively:
;
; and
.
Chapter 3. Commuting Involution Graph of 2. Sn
37
Then by Lemma 3.4.2 we have that a commutes with x, whereas a does not commute
with w or y.
3.5
The connectivity of C(2. Sn , X)
The next lemma gives a way to prove the connectivity of the commuting involution
graphs of finite groups. We will denoted by CC the connected component containing t
inside C(G, X). Also we will denote by StG (S) the stabilizer in a finite group G of a
finite set S.
Lemma 3.5.1. Let G be a finite group and X be a conjugacy class of involutions G
containing t. Then the graph C(G, X) is connected if and only if G = StG (CC ).
Proof. Set S = StG (CC ). If C(G, X) is connected then we must have CC = X. For
any g ∈ G and x ∈ CC the element xg ∈ CC so that G = S. To prove the other side
of the argument assume that G = S. If t ∈ CC , then as CC S = CC , we have tS ⊆ CC .
Here X = CC and the connectivity of the graphs is satisfied.
Theorem 3.5.2. Let X be a conjugacy class of involutions in 2. Sn . Then C(2. Sn , X)
is disconnected if and only if either n = 2m + 1, n = 6 and m = 3 or n > 4 and m = 1.
Proof. If n = 4 and m = 1 or n = 2m + 1, then C(2. Sn , X) is disconnected by
Theorem 2.3.1. Thus we may assume that n = 6 and m = 3. In this case Lemma
3.4.2, shows that there is no path between [1 2][3 4][5 6] and [1 3][2 4][5 6]. The
same reason shows that if n > 4 and m = 1, there is no path between [1 2] and [1 3]
. Therefore, we need to prove for n > 8 that C(2. Sn , X) is a connected commuting
involution graph. To prove this, since 2. Sn is generated by elements of the form [α β]
such that α, β ∈ Ω. Hence, by Lemma 3.5.1 it is satisfactory to prove a and ar are
connected for all element r of the form [α β]. Let r = [α β] where α, β ∈ Ω. Now if
α, β ∈ f ix(a) or αa = β then it is understandable that either a = ar or a = tar . In
such case clearly there is a path between a and ar . Now suppose that α, β ∈ supp(a)
and αa = β1 and β a = β2 . Let a = [1 2][3 4][5 6] · · · [α β1 ][β2 β] · · · [2m − 1 2m]
so that ar = ±[1 2][3 4][5 6] · · · [β β1 ][β2 α] · · · [2m − 1 2m]. Since n > 8, we may
3.6. The Diameter of C(2. Sn , X)
38
set w = [1 3][2 4][5 6] · · · [α β][β1 β2 ] · · · [2m − 1 2m] and then Lemma 3.4.2 gives
d(a, w) = d(w, ar ) = 1.
Finally, assume without loss of generality that α ∈ f ix(a) and β ∈
/ f ix(a) so let
β a = β1 . Hence without loss of generality let a = [1 2][3 4][5 6] · · · [β1 β] · · · [2m − 1 2m]
so that ar = ±[1 2][3 4][5 6] · · · [β1 α] · · · [2m − 1 2m]. If n > 2m + 3 set w =
[1 3][2 4][5 6] · · · [c d] · · · [2m − 1 2m] where c, d ∈ f ix(a) \ {α}. Also β1 , α ∈ f ix(w).
Thus Lemma 3.4.2 proves that d(a, w) = d(w, ar ) = 1 in this case. On the other
hand, if n = 2m + 2, then let w1 = [1 3][2 4][c α] · · · [β1 β] · · · [2m − 1 2m] and w2 =
[1 4][2 3][β1 α] · · · [c β] · · · [2m − 1 2m] such that c ∈ f ix(a) and c 6= α. Now by using
Lemma 3.4.2 we get d(a, w1 ) = d(w1 , w2 ) = d(w2 , ar ) = 1 ( note that f ix(w1 ) = {5, 6}).
This completes the proof of the theorem .
3.6
The Diameter of C(2. Sn , X)
In this section we consider the diameters of the commuting involution graphs C(2. Sn , X).
The study includes giving specific value to the restrictions of the diameters of the graph.
In fact we going to prove that the diameters of the graph is almost 4. Note that for a
group G and x, y ∈ G we denoted by [x, y] = x−1 y −1 xy the commutator of x and y
in G.
Proposition 3.6.1. Let X be a conjugacy class of involution in a double cover of Sn ,
such that n > 8 and let x ∈ X. If there is no α ∈ f ix(a) such that αx ∈
/ f ix(a), then
d(a, x) 6 3.
Proof. The aim is to find b, b′ ∈ X such that d(a, b) = d(b, b′ ) = d(b′ , x) =
1. Then we have d(a, x) 6 3. Suppose that the connected components of Gx are
C1 , C2 , C3 , . . . , Cu . As any connected component Ci of the x-graph Gx is generated by
a part from a and a part from x , it is possible to put ai and xi as a represent of the
component parts of Ci in Gx . We should note that there are no edges between white
and black vertices in the x-graph Gx . Therefore, according to the Lemma 3.2.2 we have
to consider the following cases:
Chapter 3. Commuting Involution Graph of 2. Sn
39
First, we assume that Ci is a cycle containing only black vertices, put ai = [1 2][3 4] · · ·
[2k − 1 2k] , and xi = [2 3][4 5] · · · [1 2k]. In order to find a path between ai and xi , we
have to consider the following cases which depend on k:
Case 1.1: If k = 2. In this case, we may assume without loss of generality that ai = [1 2][3 4]
and xi = [2 3][1 4]. Thus the corresponding part in Gx is given by
. We
have the following subcases:
i- if Gx has the following components
, then
we may write x as follows:
Q1
Q2
Q3
Q4
x = [2 3][1 4] ri=1
[βi βi′ ] ri=1
[λi λi + 1] ri=1
[δi ] ri=1
[αi ] , where βi , βi′ , δi ∈
Q2
Q2
f ix(a) and λi , λi + 1, αi ∈ supp(a) such that ri=1
[λi λi + 1] = ri=1
[2i +
3 2i + 4]. To find path between a and x we note the following two cases:
• If r1 is odd so that by Lemma 3.4.2 we get d(a, x) = 1 .
• If r1 even (if r1 = 0 since n > 8 take b = [2 4][1 3][λ1 λ2 ][λ1 + 1 λ2 +
Q2
Q3
Q4
Qr1 −2
1] ri=3
[λi λi + 1] ri=1
[δi ] ri=1
[αi ]), otherwise set b = [1 2][3 4] i=1
Q2
Q3
Q4
[βi βi′ ][βr1 βr1 −1 ][βr′ 1 βr′ 1 −1 ] ri=1
[λi λi + 1] ri=1
[δi ] ri=1
[αi ], thus by
Lemma 3.4.2 we obtain [a, b] = [b, x] = 1.
ii- If Gx has a connected component Cj such that Cj is a cycle of length 2,
therefore, we may assume that aj = [5 6][7 8] and xj = [6 7][8 5] be
the generator of Cj in Gx . Now put aij = ai aj = [1 2][3 4][5 6][7 8]
and xij = xi xj = [2 3][1 4][6 7][8 5]. The corresponding part in Gx is
, so that Lemma 3.4.2 gives d(aij , xij ) = 1.
iii- If Gx has a connected component Cj such that Cj is a cycle of length 3
,
, then let aj = [5 6][7 8][9 10] and xj = [6 7][8 9][10 5]. Here
put aij = [1 2][3 4][5 6][7 8][9 10] and xij = [2 3][1 4][6 7][8 9][10 5].
a
b
In this case set bij = [1 3][2 4][5 6][7 10][8 9]. Thus Gbijij ,Gxijij are given
by
d(bij , xij ) = 1.
and by Lemma 3.4.2, we obtain d(aij , bij ) =
3.6. The Diameter of C(2. Sn , X)
40
iv- If Gx has a connected component Cj such that Cj is a chain of length
one consisting of only black vertices
, then let aj = [5 6][7 8] and
xj = [6 7][5][8]. In this case Gx has a connected component Cl such that
al = [β1 ][β2 ] and xl = [β1 β2 ] where β1 , β2 ∈ f ix(a), so that if aijl =
[1 2][3 4][5 6][7 8][β1 ][β2 ] and xijl = [2 3][1 4][6 7][β1 β2 ]. taking bijl =
a
b
ijl
[1 3][2 4][5 8][6 7][β1 ][β2 ][5][8], we have that Gbijl
,Gxijl
ijl are given respectively
by
and
. Hence aijl
and bijl commute, as do bijl and xijl . Thus d(aijl , bijl ) = d(bijl , xijl ) = 1.
v- If Gx has a connected component Cj such that Cj is a cycle, then, let aj =
[5 6][7 8] · · · [2r + 3 2r + 4] and xj = [6 7][8 9] · · · [2r + 2 2r + 3][5 2r + 4]
where r > 3. Now put aij = [1 2][3 4][5 6][7 8] · · · [2r + 3 2r + 4] and
a
xij = [2 3][1 4][6 7][8 9] · · · [2r + 2 2r + 3][5 2r + 4]. It follows that Gxijij is
given by
. Therefore, our aim is to find bij
a
b
b
b
and b′ij (if required) corresponding to aij , xij such that Gbijij , Gxijij , Gb ij′ , Gxijij
′
ij
are given by,
and the total number of
is even.
Thus Lemma 3.4.2 gives a path between aij , bij , b′ij and xij . However, in
order to do that we need to take into account the following cases:
• If r is even such that r = 2s and s is even, set bij = [2 3][1 4][5 6][7 2r +
4][8 2r +3] · · · [r +5 r +6]. In this case the number of
a
b
in Gbijij , Gxijij
is s .
• If r is even such that r = 2s and s is odd, put bij = [1 2][3 4][5 6][7 2r +
a
b
4][8 2r + 3] · · · [r + 5 r + 6]. We see that in Gbijij , Gxijij the total number
of
respectively equal to s − 1, s + 1 .
• If r is odd such that r = 2s+1 and s is odd, then let bij = [1 3][2 4][5 6][7
2r + 4][8 2r + 3] · · · [r + 5 r + 6]. The total number of
a
b
in Gbijij , Gxijij
is s + 1 .
• If r = 2s + 1 and s is even, then let bij = [1 2][3 4][5 6][7 2r + 4][8 2r +
Chapter 3. Commuting Involution Graph of 2. Sn
41
3] · · · [r+5 r+6], and b′ij = [2 3][1 4][5 7][6 2r+4][8 2r+3] · · · [r+5 r+6].
b
b
′
b
ij
ij
in Gaij
ij , Gxij equal to s while in Gb ′ it equals
Then the number of
ij
2.
vi- If Gx has a connected component Cj such that Cj is a chain, then let aj =
[5 6][7 8] · · · [2r + 3 2r + 4] and xj = [6 7][8 9] · · · [2r + 2 2r + 3]. In this case
Gx has a connected component Cl such that al = [β1 ][β2 ], and xl = [β1 β2 ]
where β1 , β2 ∈ f ix(a) and r > 2. Thus let aijl = [1 2][3 4][5 6][7 8] · · · [2r +
3 2r+4][β1 ][β2 ] and xijl = [2 3][1 4][6 7][8 9] · · · [2r+2 2r+3][β1 β2 ][5][2r+4].
a
ijl
is given by
Then Gxijl
. Hence,
our target is to find bijl and b′ijl (if required) corresponding to aijl , xijl such
a
b
b
b
′
ijl
ijl
ijl
that Gbijl
, Gxijl
′ , Gxijl are given by,
ijl , Gb
ijl
and the total number of
and
, is both even so that by Lemma 3.4.2 we may find a
path between the corresponding parts aijl , bijl , b′ijl ,and xijl . To deal with
that we to consider the following cases:
• If r is even such that r = 2s and s is even, then set bijl = [1 2][3 4][5 2r+
3][6 2r + 4][7 2r + 2] · · · [r + 4 r + 5], and b′ijl = [2 3][1 4][5 2r + 4][6 2r +
3][7 2r + 2][8 2r + 1] · · · [r + 4 r + 5]. In this case the total number of
,
a
b
b
ijl
in Gbijl
respectively s, 0 and in Gb ijl′ 2, 0. Moreover, in Gxijl
ijl
′
ijl
the total number equals s − 1, 1.
• If r is even such that r = 2s and s is odd, then let bijl = [1 3][2 4][5 2r +
4][6 2r + 3] · · · [r + 4 r + 5]. The total number of
,
b
a
ijl
in Gbijl
respectively equals s + 1, 0,and in Gxijl
ijl the total number equals s, 1.
• If r is odd such r = 2s + 1 and s is odd, then let bijl = [2 3][1 4][5 2r +
a
ijl
4][6 2r + 3] · · · [r + 4 r + 5], so that in Gbijl
the total number of
,
b
equal s + 1, 0, while in Gxijl
ijl it equals s, 1.
• If r = 2s + 1 and s is even, set bijl = [1 2][3 4][5 2r + 4][6 2r + 3] · · · [r +
4 r + 5]. In this case the total number of
b
in Gxijl
ijl is s + 1, 1.
,
a
ijl
in Gbijl
is s, 0, and
3.6. The Diameter of C(2. Sn , X)
42
Case 1.2: If k = 3, then we may assume without loss of generality that ai = [1 2][3 4][5 6]
and xi = [2 3][4 5][1 6], and Gxaii is given by
. We need to consider the
following subcases:
i- if Gx has the following components
, then
we may write x as follows:
Qr1
i=1 [βi
βi′ ]
Q r2
Q r3
Q r4
where βi , βi′ , δi
Q2
Q2
∈ f ix(a) and λi , λi + 1, αi ∈ supp(a) such that ri=1
[λi λi + 1] = ri=1
[2i +
x = [2 3][4 5][1 6]
i=1 [λi
λi +1]
i=1 [δi ]
i=1 [αi ],
5 2i + 6]. To find a path between a and x we note the following two cases:
• If r1 is odd, set b = [1 2][4 5][3 6]
Qr1
i=1 [βi
Q r1
βi′ ]
Q r2
i=1 [λi
Q r2
λi +1]
Q r3
i=1 [δi ]
Q r4
Q r3
i=1
[αi ], and let b′ = [1 3][2 6][4 β1 ][5 β1′ ] i=2 [βi βi′ ] i=1 [λi λi + 1] i=1 [δi ]
Qr4
′
′
i=1 [αi ]. Hence by Lemma 3.4.2, we obtain d(a, b) = d(a, b ) = d(x, b ) =
1.
• If r1 is even, (if r1 = 0 then as n > 8 we may set b = [1 λ1 +
Q2
Q3
Q4
1][2 λ1 ][3 5][4 6] ri=2
[λi λi +1] ri=1
[δi ] ri=1
[αi ], and b′ = [1 3][2 6][4 λ1 ]
Q2
Q3
Q4
[5 λ1 + 1] ri=2
[λi λi + 1] ri=1
[δi ] ri=1
[αi ]), otherwise we put b =
Q3
Q4
Q1
Q2
[1 2][3 4][5 6] ri=1
[βi βi′ ] ri=1
[λi λi + 1] ri=1
[δi ] ri=1
[αi ], and b′ =
Q2
Q3
Q4
Q1
[βi βi′ ][β1 β2 ][β1′ β2′ ] ri=1
[λi λi + 1] ri=1
[δi ] ri=1
[αi ].
[1 2][4 5][3 6] ri=2
Again we have by Lemma 3.4.2, that d(a, b′ ) = d(b′ , b) = d(x, b′ ) = 1.
ii- If Gx has a connected component Cj such that Cj is a cycle of length 2,
similar to the above case 1.1(iii) there is path between ai and xi .
iii- If Gx has a connected component Cj such that Cj is a cycle of length 3,
we may let aj = [7 8][9 10][11 12] and xj = [8 9][10 11][7 12]. Here put
aij = [1 2][3 4][5 6][7 8][9 10][11 12] and xij = [2 3][4 5][1 6][8 9][10 11][7 12].
a
b
In this case put bij = [1 2][4 5][3 6][7 8][10 11][9 12], so that Gbijij ,Gxijij are
given by
d(aij , bij ) = d(bij , xij ) = 1.
. Hence by Lemma 3.4.2, we get
Chapter 3. Commuting Involution Graph of 2. Sn
43
iv- If Gx has a connected component Cj such that Cj is a chain of length one
, then let aj = [7 8][9 10] and
consisting of only black vertices
xj = [8 9]. In this case Gx has a connected component Cl such that,
al = [β1 ][β2 ] and xl = [β1 β2 ] where β1 , β2 ∈ f ix(a). If we let aijl =
[1 2][3 4][5 6][7 8][9 10][β1 ][β2 ], and xijl = [2 3][4 5][1 6][8 9][β1 β2 ][7][10],
a
b
ijl
also set bijl = [1 2][4 5][3 6][8 9][7 10][β1 ][β2 ], then Gbijl
,Gxijl
ijl are given
respectively by
and
.
Therefore, Lemma 3.4.2 gives d(aijl , bijl ) = d(bijl , xijl ) = 1.
v- If Gx has a cconnected component Cj such that Cj is a cycle, then, let aj =
[7 8][9 10] · · · [2r + 5 2r + 6] and xj = [8 9][10 11] · · · [2r + 4 2r + 5][7 2r + 6]
where r > 3. Now put aij = [1 2][3 4][5 6][7 8] · · · [2r + 5 2r + 6] and
a
xij = [2 3][4 5][1 6][8 9][10 11] · · · [2r + 4 2r + 5][7 2r + 6]. Then Gxijij is given
. Therefore, our aim is to find bij and b′ij
by
a
b
b
′
b
(if required) corresponding to aij , xij such that Gbijij , Gxijij , Gb ij′ , Gxijij are given
ij
by,
and the total number of
is even.
Thus Lemma 3.4.2 gives a path between aij , bij , b′ij and xij . To do that we
need to take into account the following cases:
• If r = 2s with s is even, then set bij = [1 2][3 4][5 6][7 2r + 6][8 2r +
5][9 2r + 4]...[r + 5 r + 8][r + 6 r + 7], and b′ij = [1 2][4 5][3 6][7 2r +
6][8 2r + 4][9 2r + 5][10 2r + 3][11 2r + 2] · · · [r + 6 r + 7]. In this case the
a
b
′
b
in Gbijij , Gxijij is equal to s, while it is 2 in Gb ij′ .
total number of
ij
• If r = 2s with s is odd, then set bij = [1 2][4 5][3 6][7 2r + 6][8 2r +
5][9 2r + 4] · · · [r + 5 r + 8][r + 6 r + 7], and b′ij = [1 3][4 5][2 6][7 r +
6][2r + 6 r + 7][8 2r + 5][9 2r + 4] · · · [r + 5 r + 8]. The total number of
a
b
′
b
in Gbijij , Gxijij it is equal to s + 1, and is equal 2 in Gb ij′ .
ij
• If r = 2s + 1 with s is even, then set bij = [1 2][3 4][5 6][7 2r + 6][8 2r +
5][9 2r + 4] · · · [r + 5 r + 8][r + 6 r + 7], and b′ij = [1 2][4 5][3 6][7 2r +
3.6. The Diameter of C(2. Sn , X)
44
6][8 2r + 5] · · · [r + 4 r + 9][r + 5 r + 6][r + 7 r + 8]. The total number of
a
b
′
b
in Gbijij , Gxijij is equal to s, and it is 2 in Gb ij′ .
ij
• If r = 2s + 1 with s is odd, then set bij = [1 2][4 5][3 6][7 2r + 6][8 2r +
5][9 2r + 4] · · · [r + 5 r + 8][r + 6 r + 7]. In this case the total number of
a
b
in Gbijij , Gxijij is s + 1.
vi- If Gx has connected component Cj such that Cj is a chain, then let aj =
[7 8][9 10] · · · [2r + 5
2r + 6] and xj = [8 9][10 11] · · · [2r + 4 2r + 5].
In this case Gx has a connected component Cl such that al = [β1 ][β2 ],
β2 ] where β1 , β2 ∈ f ix(a) and r > 2. Thus let aijl =
and xl = [β1
[1 2][3 4][5 6][7 8][9 10] · · · [2r + 5 2r + 6][β1 ][β2 ] and xijl = [2 3][4 5][1 6][8 9]
a
ijl
[10 11] · · · [2r + 4 2r + 5][β1 β2 ]. Then Gxijl
is given by
. Hence, our target is to find bijl
a
b
b
b
ijl
ijl
ijl
and b′ijl (if required) corresponding to aijl , xijl such that Gbijl
, Gxijl
′ , Gxijl
ijl , Gb
′
ijl
are given by,
and the total numand
ber of
is even. Then by Lemma 3.4.2 we get paths between
aijl , bijl , b′ijl and xijl . For that we consider the following cases:
• If r = 2s and s is even, then set bijl = [1 2][3 4][5 6][7 2r + 6][8 2r +
5] · · · [r + 5 r + 8][r + 6 r + 7] and b′ijl = [1 2][3 6][4 5][β1 β2 ][8 2r +
5] · · · [r + 5 r + 8][r + 6 r + 7]. In this case the total number of
a
b
′
,
b
ijl
ijl
in Gbijl
, Gxijl
′ the numbers
ijl respectively equals s, 0. Moreover in Gb
ijl
equal {1, 1}.
• If r = 2s and s is odd, then set bijl = [1 2][3 6][4 5][7 2r + 6][8 2r +
5] · · · [r + 5 r + 8][r + 6 r + 7]. It follows the total number of
a
,
b
ijl
in Gbijl
, Gxijl
ijl respectively {s + 1, 0} and {s, 1}.
• If r = 2s + 1 with s is even, then take bijl = [1 2][3 6][4 5][7 2r +
6][8 2r + 5] · · · [r + 5 r + 8][β1 β2 ] and b′ijl = [2 3][1 6][4 5][β1 β2 ][8 2r +
5] · · · [r + 5 r + 8][r + 6 r + 7]. Hence the total number of
,
in
Chapter 3. Commuting Involution Graph of 2. Sn
a
b
45
′
b
ijl
Gbijl
, Gb ijl′ , Gxijl
ijl respectively equals {s + 1, 1}, {1, 1} and {s, 0}.
ijl
• If r = 2s+1 with s is odd, let bijl = [1 2][3 6][4 5][7 2r+6][8 2r+5] · · · [r+
5 r+8][r+6 r+7] and b′ijl = [1 3][2 6][4 5][β1 β2 ][8 2r+5] · · · [r+6 r+7].
Thus the total number of
b
,
a
b
′
ijl
in Gbijl
, Gxijl
ijl is {s + 1, 0}, and in
Gb ijl′ is {1, 1}.
ijl
Case 1.3: If k > 3, then Gxaii is given by
. Therefore, our aim is to find bi
and b′i if required corresponding to ai , xi such that Gbaii , Gxbii , Gbbi′ , Gxbii are given by
′
i
and the total number of
is even. Thus
Lemma 3.4.2 gives paths between ai , bi , b′i and xi . We need to take into account
the following cases:
i- If k = 2s and s is even, then bi = [1 k + 1][2 k + 2][3 k + 3][4 k + 4] · · · [k 2k].
In this case the total number of
in Gbaii , Gxbii is equal to s.
ii- If k = 2s and s is odd, then set bi = [1 2][4 2k][3 2k − 1][5 2k − 2] · · · [k −
1 k + 4][k k + 3][k + 1 k + 2] and b′i = [1 2][3 2k][4 2k − 1] · · · [k − 1 k +
4][k k + 1][k + 2 k + 3]. Here the total number of
in Gbaii , Gxbii is equal
′
to s − 1, and in Gbbi′ 2.
i
iii- If k = 2s + 1 and s is odd, then set bi = [1 2][3 4][2k − 1 2k][5 2k −
2][6 2k − 3] · · · [k − 2 k + 5][k − 1 k + 4][k k + 3][k + 1 k + 2] and b′i =
[1 2][3 2k][4 2k − 1][5 2k − 2] · · · [k − 2 k + 5][k − 1 k][k + 3 k + 4][k + 1 k + 2].
Then the total number of
in Gbaii , Gxbii is equal to s − 1, while in Gbbi′
′
i
2.
iv- If k = 2s + 1 and s is even, then put bi = [1 2][3 2k][4 2k − 1] · · · [k + 1 k + 2],
so that the total number of
in Gbaii , Gxbii is s.
Secondly, if Ci is a chain containing only black vertices, then we may write ai =
[1 2][3 4] · · · [2k − 1 2k] , and xi = [2 3][4 5] · · · [2k − 2 2k − 1]. We note that Gx has a
connected component Cl such that al = [β1 ][β2 ] and xl = [β1 β2 ], where β1 , β2 ∈ f ix(a).
Thus let ail = [1 2][3 4] · · · [2k −1 2k][β1 ][β2 ] and xil = [2 3][4 5] · · · [2k −2 2k −1][β1 β2 ].
3.6. The Diameter of C(2. Sn , X)
46
It follow that Gxailil is given by
, so that in order to find
a path between ail and xil , we have to consider the following cases:
Case 2.1: If k = 2, we may assume that ail = [1 2][3 4][β1 ][β2 ] and xil = [2 3][β1 β2 ]. In this
case Gxailil is
, so that to find a path between them, we need to
argue the following cases:
i- if x = [2 3]
Qr1
i=1 [µi
µ′i ]
Qr2
i=1 [λi
λi + 1]
Q r3
i=1 [δi ]
Qr4
where λi , λi + 1, δi
Qr2
i=1 [λi λi +1] =
i=1 [2i+3 2i+
Q r2
i=1 [αi ],
∈ supp(a) and αi , µi , µ′i ∈ f ix(a) such that
Q3
Q2(r1 +r2 )+2
4] and ri=1
[δi ] = [1][4] i=2r
[i] if r1 6= 1. Without loss of generality
2 +5
we may let [δ1 ] = [1], [δ2 ] = [4] and [β1 β2 ] = [µ1 µ′1 ]. Then Gx has the
following components
. Thus
to find path between a and x we need to consider the following sub cases:
• if r1 is odd (if r1 = 1 then since n > 8 we let b = [2 3][1 4][λ1 λ2 ][λ1 +
Q2
Q4
Q1
1 λ2 +1] ri=2
[λi λi +1] ri=1
[αi ][β1 ][β2 ] ), we set b = [2 3][1 4][δ3 δ4 ] ri=3
Q2
Q3
Q4
[µi µ′i ] ri=1
[λi λi + 1] ri=5
[δi ] ri=1
[αi ][µ1 ][µ′1 ][µ2 ][µ′2 ], where δ3 , δ4 ∈
(supp(a)∩f ix(x))\{1, 4}. Thus Lemma 3.4.2 gives d(a, b) = d(b, x) = 1.
Q2
Q1
[λi λi +
• if r1 is even then put b = [δ3 δ4 ] ri=3
[µi µ′i ][µ1 µ2 ][µ′1 µ′2 ] ri=1
Q4
Q3
1] ri=5
[δi ][1][4][2][3] ri=1
[αi ]. Again δ3 , δ4 ∈ (supp(a) ∩ f ix(x)) \ {1, 4}
, so d(a, b) = d(b, x) = 1 by Lemma 3.4.2.
ii- If Gx has a connected component Cj such that Cj is a cycle of length 2, we
may assume that aj = [5 6][7 8] and xj = [6 7][8 5]. Put ailj = ai al aj =
[1 2][3 4][5 6][7 8][β1 ]β2 ] and xilj = xi xl xj = [2 3][β1 β2 ][6 7][8 5]. Now
a
b
ilj
let bilj = [2 3][1 4][5 7][6 8], so that Gbilj
,Gxilj
ilj are given respectively by
and
. Hence by Lemma 3.4.2 one
can see that d(ailj , bilj ) = d(bilj , xilj ) = 1.
iii- If Gx has a connected component Cj such that Cj is a cycle of length 3
, then let aj = [5 6][7 8][9 10] and xj = [6 7][8 9][10 5]. Here put
ailj = [1 2][3 4][5 6][7 8][9 10][β1 ][β2 ] and xilj = [2 3][β1 β2 ][6 7][8 9][10 5].
Chapter 3. Commuting Involution Graph of 2. Sn
47
a
b
ilj
Let bilj = [2 3][1 4][5 6][7 10][8 9], so that Gbilj
,Gxilj
ilj are given respectively
by
and
. Therefore, Lemma 3.4.2 demonstrates that
d(ailj , bilj ) = d(bilj , xilj ) = 1.
iv- If Gx has a connected component Cj such that Cj is a chain
, then
let aj = [5 6][7 8] and xj = [6 7]. Similar to above we can find a connected component Cq such that aq = [β3 ][β4 ] and xq = [β3
β4 ] where
β3 , β4 ∈ f ix(a). Thus if ailjq = [1 2][3 4][5 6][7 8][β1 ][β2 ][β3 ][β4 ] and xiljq =
a
b
iljq
[2 3][β1 β2 ][6 7][β3 β4 ], and we let biljq = [2 3][1 4][5 8][6 7], then Gbiljq
,Gxiljq
iljq
are given respectively by
and
. Con-
sequently using Lemma 3.4.2 we have d(ailjq , biljq ) = d(biljq , xiljq ) = 1.
v- If Gx has a connected component Cj such that Cj is a cycle with aj =
[5 6][7 8] · · · [2r + 3
2r + 4] and xj = [6 7][8 9] · · · [5
2r + 4] where
r > 3, then let ailj = [1 2][3 4][5 6][7 8] · · · [2r + 3 2r + 4] and xilj =
a
ilj
[2 3][β1 β2 ][6 7][8 9] · · · [2r + 2 2r + 3][5 2r + 4]. Here Gxilj
is given by
. Hence, our target is to find bilj and b′ilj (if
a
b
b
′
b
ilj
ilj
ilj
required) corresponding to ailj , xilj such that Gbilj
, Gxilj
′ , Gxilj are given
ilj , Gb
ilj
by,
and the total number of
and
is even. Hence Lemma 3.4.2 gives edges between
aijl , bijl , b′ijl ,and xijl . Therefore, to deal with above we need the following
cases:
• If r is even such that r = 2s and s is even, then set bilj = [1 3][2 4][5 6][7 2r+
4][8 2r+3] · · · [r+5 r+6] and b′ilj = [1 4][2 3][6 7][5 2r+4][8 2r+3] · · · [r+
a
5 r + 6]. In this case the total number of
b
b
′
ilj
, in Gbilj
, Gxilj
ilj respec-
tively equals {s, 0},{s − 1, 1}. Moreover, in Gb ilj′ equals 2, 0.
ilj
• If r is even such that r = 2s and s is odd, then set bilj = [1 3][2 4][5 r +
5][6 r + 6][7 2r + 4][8 2r + 3] · · · [r + 3 r + 8][r + 4 r + 7] and b′ilj =
3.6. The Diameter of C(2. Sn , X)
48
[1 4][2 3][5 6][7 2r + 4][8 2r + 3] · · · [r + 5 r + 6]. Consequntly the total
,
number of
b
a
b
′
ilj
in Gbilj
, Gxilj
ilj respectively equals {s + 1, 0},{s, 1},
while in Gb ilj′ is 2, 0.
ilj
• If r is odd such that r = 2s+1 and s is odd, put bilj = [1 4][2 3][5 6][7 2r+
4][8 2r + 3] · · · [r + 5 r + 6]. Here the total number of
a
,
in
b
ilj
Gbilj
, Gxilj
ilj respectively {s + 1, 0},{s, 1}.
• If r = 2s + 1 and s is even, then let bilj = [1 4][2 3][5 6][7 8][2r + 3 2r +
4][9 2r + 2] · · · [r + 5 r + 6] and b′ilj = [1 4][2 r + 5][3 r + 6][5 6][7 2r +
a
b
ilj
, in Gbilj
, Gxilj
ilj
4][8 2r + 3] · · · [r + 4 r + 7]. The total number of
b
′
respectively {s, 0},{s + 1, 1}, and in Gb ilj′ 2, 0.
ilj
vi- If Gx has a connected component Cj such that Cj is a chain with aj =
[5 6][7 8] · · · [2r + 3 2r + 4]
and
xj = [6 7][8 9] · · · [2r + 2
2r + 3],
then the x-graph Gx has a connected component Cq such that aq = [β3 ][β4 ]
and xq = [β3
β4 ] where β3 , β4 ∈ f ix(a) and r > 2. Now let ailjq =
[1 2][3 4][5 6][7 8] · · · [2r+3 2r+4][β1 ][β2 ][β3 ][β4 ] and xiljq = [2 3][β1 β2 ][6 7]
a
iljq
is given by
[8 9] · · · [2r + 2 2r + 3][β3 β4 ]. The graph Gxiljq
.
Hence, our target is to find biljq and b′iljq (if required) corresponding to
a
b
b
′
b
iljq
iljq
iljq
ailjq , xiljq such that Gbiljq
, Gxiljq
′ , Gxiljq are given by,
iljq , Gb
iljq
and the total number of
and
is even. Hence Lemma 3.4.2 gives paths between
ailjq , biljq , b′iljq and xiljq . We need to take into account the following cases:
• If r is even such that r = 2s and s is even, then set biljq = [1 2][3 4][5 2r+
4][6 2r + 3][7 2r + 2] · · · [r + 2 r + 7][r + 3 r + 6][r + 4 r + 5], and
b′iljq = [2 3][1 4][5 2r +4] · · · [r +1 r +8][r +2 r +3][r +6 r +7][r +4 r +5].
In this case the total number of
,
b
a
b
′
iljq
in Gbiljq
, Gxiljq
iljq respectively
equals {s, 0},{s − 2, 2}. Moreover in Gb iljq′ it equals 2, 0.
iljq
• If r is even such that r = 2s and s is odd, then let biljq = [1 4][2 3][5 2r+
4][6 2r + 3] · · · [r + 4 r + 5], so that the total number of
,
in
Chapter 3. Commuting Involution Graph of 2. Sn
a
49
b
iljq
Gbiljq
, Gxiljq
iljq respectively {s + 1, 0},{s − 1, 2}.
• If r is odd such r = 2s + 1 and s is odd, then put biljq = [1 4][2 3][5 2r +
4][6 2r + 3][7 2r + 2] · · · [r + 2 r + 7][r + 3 r + 6][r + 4 r + 5] and
b′iljq = [β1 β2 ][β3 β4 ][5 2r + 4][6 2r + 3][7 2r + 2] · · · [r + 2 r + 7][r +
a
b
iljq
, in Gbiljq
, Gxiljq
iljq
3 r + 6][r + 4 r + 5]. Then the total number of
b
′
respectively {s + 1, 0},{s, 1}, and in Gb iljq′ 0, 2.
iljq
• If r = 2s + 1 and s is even, let biljq = [1 2][3 4][5 2r + 3][6 2r +
4][7 2r + 2] · · · [r + 4 r + 5] and b′iljq = [1 4][2 3][5 2r + 4][6 2r +
3][7 2r + 2][8 2r + 1] · · · [r + 4 r + 5]. The total number of
a
b
′
,
in
b
iljq
iljq
Gbiljq
, Gxiljq
′ the total numbers
iljq respectively {s, 0},{s, 2}. Also in Gb
iljq
is 2, 0.
Case 2.2: If k > 2, our target is to find bil and b′il (if required) corresponding to ail , xil such
that Gbailil , Gxbilil , Gbbil′ , Gxbilil are given by,
′
il
and the total number of
and
is even. Lemma 3.4.2 then gives a path between ail , bil , b′il and xil .
However, to cope with that we need to consider the following cases:
i- If k = 2s and s is even, then set bil = [1 2k][2 2k − 1][3 2k − 2] · · · [k k + 1].
In this case the total number of
,
in Gbailil , Gxbilil respectively equals
{s, 0},{s − 1, 1}.
ii- If k = 2s and s is odd, then let bil = [1 2][2k − 1 2k][3 2k − 2][4 2k − 3] · · · [k −
2 k + 3][k − 1 k + 2][k k + 1] and b′il = [1 2k][2 2k − 1][3 2k − 2] · · · [k − 3 k +
4][k − 2 k − 1][k + 2 k + 3][k k + 1]. Hence the total number of
,
in Gbailil , Gxbilil respectively {s − 1, 0},{s − 2, 1}. Moreover in Gbbil′ the total
′
il
numbers 2, 0.
iii- If k = 2s + 1 and s is odd, then if k = 3 let bil = [1 5][2 6][β1 β2 ] and
b′il = [1 6][2 5][3 4]. Otherwise set bil = [1 2k − 1][2 2k][3 2k − 2][4 2k −
3][5 2k − 4] · · · [k − 1
k + 2][β1
β2 ] and b′il = [1 2k][2 2k − 1][3 2k −
2][4 2r − 3][5 2k − 4] · · · [k k + 1]. The total number of
, in Gbailil , Gxbilil
′
3.6. The Diameter of C(2. Sn , X)
50
respectively {s, 1}, whilst in Gbbil′ equals 1, 1.
il
iv- If k = 2s + 1 and s is even, then let bil = [1 2k − 1][2 2k][3 2k − 2] · · · [k −
1 k + 2][k k + 1] and b′il = [1 2k][2 2k − 1][3 2k − 2] · · · [k − 2 k + 3][k − 1 k][k +
, in Gbailil , Gxbilil respectively equals
′
1 k + 2]. Here the total number of
{s, 0},{s − 1, 1}. Additionally in Gbbil′ the number of such components is 2, 0.
il
Finally, If Ci has non-conected black vertices, so that Gxaii is given by following
, and ai and xi are the corresponding parts of Ci , we may let ai =
Q1
[1 2][3 4] · · · [2k − 1 2k] and xi = ri=1
[βi βi′ ], where βi , βi′ , ∈ f ix(a). To find a path
between ai and xi we consider the following cases:
Case 3.1: If r1 is even, then by Lemma 3.4.2 we get d(ai , xi ) = 1.
Case 3.2: If r1 6= 1 is odd, then set bi = [1 3][2, 4]
of
,
Qr1 −2
i=1
[βi βi′ ]. In this case the total number
in Gbaii , Gxbii respectively equals {1, r1 − 2},{0, 2}.
Case 3.3: If r1 = 1, we may assume that ai = [1, 2] and xi = [2m + 1 2m + 2]. Here we only
need to care about the following subcases:
1: If Gx has a connected component Cj such that Cj is a cycle, let aj =
[3 4][5 6] · · · [2r + 1 2r + 2] and xj = [4 5][6 7] · · · [2r 2r + 1][3 2r + 2]
where r > 3. Now put aij = [1 2][3 4][5 6][7 8] · · · [2r + 1 2r + 2] and
a
xij = [2m + 1 2m + 2][4 5][6 7][8 9] · · · [2r 2r + 1][3 2r + 2]. The graph Gxijij is
. Hence, our target is to find bij and b′ij if
given by
a
b
b
b
′
required corresponding to aij , xij such that Gbijij , Gxijij , Gb ij′ , Gxijij are given by,
ij
and the total number of
and
is even.
Thus Lemma 3.4.2 gives edges between
aij , bij , b′ij and xij . To do that we need to take into account the following
cases:
i- If r is even such that r = 2s and s is even, set bij = [2m + 1 2m +
2][3 4][5 2r + 2][6 2r + 1][7 2r] · · · [r + 3 r + 4]. In this case the total
Chapter 3. Commuting Involution Graph of 2. Sn
,
number of
51
b
a
in Gbijij , Gxijij respectively {s − 1, 1},{s, 0}.
ii- If r is even such that r = 2s and s is odd, put bij = [1 2][3 4][5 2r +
2][6 2r + 1][7 2r] · · · [r + 3 r + 4]. Thus the total number of
a
, in
b
Gbijij , Gxijij respectively {s − 1, 0},{s, 1}.
iii- If r is odd such that r = 2s + 1 and s is odd, let bij = [2m + 1 2m +
2][3 2r + 2][4 2r + 1][5 2r] · · · [r + 2 r + 3], and bij ′ = [2m + 1 3][2m +
2 2r + 2][4 2r][2r + 1 5][6 2r − 1] · · · [r + 2 r + 3]. Thus the total number
of
,
a
b
′
b
in Gbijij , Gxijij , Gb ij′ respectively {s, 1},{s + 1, 0},{2, 0} .
ij
iv- If r = 2s+1 and s is even, then let bij = [1 2][3 4][5 2r+2][6 2r+1] · · · [r+
3 r + 4] and bij ′ = [2m + 1 2m + 2][3 5][4 2r + 2][6 2r + 1] · · · [r + 3 r + 4].
,
The total number of
a
b
′
b
in Gbijij , Gxijij is {s, 0}, whilst in Gb ij′ is
ij
{1, 1} .
2: If Gx has a connected component Cj such that Cj is a chain, then let
aj = [3 4][5 6] · · · [2r + 1 2r + 2]
and xj = [4 5][6 7] · · · [2r 2r + 1]
where r > 2. The x-graph Gx has a connected component Cq such that
aq = [β1 ][β2 ] and xq = [β1
β2 ] where β1 , β2 ∈ f ix(a). Now let aijq =
[1 2][3 4][5 6][7 8] · · · [2r + 1 2r + 2][2m + 1][2m + 2][β1 ][β2 ] and xijq =
a
ijq
is given by
[2m + 1 2m + 2][β1 β2 ][4 5][6 7][8 9] · · · [2r 2r + 1]. Then Gxijq
. Hence, our target is to find bijq and b′ijq (if rea
b
b
b
′
ijq
ijq
ijq
quired) corresponding to aijq , xijq such that Gbijq
, Gxijq
′ , Gxijq are given
ijq , Gb
ijq
by,
and the total number of
and
is even. Therefore Lemma 3.4.2 gives paths between
aijq , bijq , b′ijq and xijq .Thus to deal with that we need to check the following
cases:
i- If r is even such that r = 2s and s is even, set bijq = [1 2][3 2r +
2][4 2r + 1][5 2r] · · · [r + 2 r + 3] and bijq ′ = [1 3][2 2r + 2][4 5][2r 2r +
1][6 2r − 1] · · · [r + 2 r + 3]. In this case the total number of
,
3.6. The Diameter of C(2. Sn , X)
b
a
′
52
b
ijq
ijq
, Gxijq
in Gbijq
′ respectively equals {s, 0},{s − 2, 2},{2, 0}.
ijq , Gb
ijq
ii- If r is even such that r = 2s and s is odd, put bijq = [2m + 1 2m +
2][3 2r + 2][4 2r + 1][5 2r] · · · [r + 2 r + 3] and bijq ′ = [2m + 1 r + 3][2m +
2 r + 2][3 2r + 2][[5 2r + 1][4 2r][6 r + 5] · · · [r + 1 r + 4]. Thus the total
a
b
′
b
ijq
ijq
, Gxijq
in Gbijq
′ is {2, 0} .
ijq is {s, 1} and in Gb
,
number of
ijq
iii- If r is odd such that r = 2s + 1 and s is odd, let bij = [2m + 1 2m +
2][3 2r + 2][4 2r + 1][5 2r] · · · [r + 2 r + 3], so that the total number of
,
a
b
ijq
in Gbijq
, Gxijq
ijq is {s, 1}.
iv- If r = 2s + 1 and s is even, then let bij = [1 2][3 2r + 2][4 2r +
1][5 2r] · · · [r + 2 r + 3]. Therefore the total number of
a
,
in
b
ijq
, Gxijq
Gbijq
ijq respectively equals {s, 0},{s, 2}.
Now, let x ∈ X and suppose that Λ1 , Λ2 , Λ3 , Λ4 , are the number of the double
edges
, cycles with three edges
and unconnected black vertices
, chains with one edges of black vertices
, respectively. Then in order to find a path
between the components of a and x we need to consider first the following :
Case 4.1: If Λ1 + Λ2 + Λ3 + Λ4 equal an even number. To find a path between a and x,
such that d(x, a) = 1 we have the following subcases:
i- If Λ1 + Λ2 + Λ3 + Λ4 = 2. Then by Case 1.1, Case 1.2, Case 2.1, Case 3.1
and Case 3.2 there is a path between a and x.
ii- If Λ1 + Λ2 + Λ3 + Λ4 = r ≥ 4, such that r is even , then r = r1 + 2 where r1 is
an even number. Thus by the above we get a path between the components
of a and x.
Case 4.2: If Λ1 + Λ2 + Λ3 + Λ4 equal an odd number. To find a path between a and x, such
that d(x, a) = 1 we have the following subcases:
i- If Λ1 + Λ2 + Λ3 + Λ4 = 1, then by above results one can easily find a path
between a and x, such that d(x, a) = 1.
Chapter 3. Commuting Involution Graph of 2. Sn
53
ii- If Λ1 + Λ2 + Λ3 + Λ4 = 3 and if three of the Λi′ s equal zero. Then without
loss of generality we may assume Λ2 = Λ3 = Λ4 = 0, and Λ1 = 3 (if one of
the Λi′ s = 3 for i ∈ {2, 3, 4} then similar considerations can apply to show
that d(x, a) = 1). Let a1 = [1 2][3 4], a2 = [5 6][7 8],a3 = [9 10][11 12]
and x1 = [2 3][1 4], x2 = [6 7][5 8], x3 = [10 11][9 12]. Set a123 =
[1 2][3 4][5 6][7 8][9 10][11 12] and x123 = [2 3][1 4][6 7][5 8][10 11][9 12],
123 is given by
so that the x-graph Gxa123
. Let
b123 = [1 2][3 4][6 8][5 7][10 11][9 12]. In this case the total number of
123
, in Gba123
, Gxb123
equals 2.
123
iii- If two or one of Λi′ s equal zero then by the same argument as in [ii] we can
find a path in this case.
iv- If Λ1 + Λ2 + Λ3 + Λ4 = r > 3, such that r is odd, then r = r1 + 3, where
r1 is an even number and again by the above we have a path between the
components of a and x, such that d(a, x) = 1.
Finally, let b and b′ be the product of all the bi′ s and bi′ ‘s , as mentioned above. It is
not hard to see that |f ix(a)| = f ix(b)|=|f ix(b′ )|. Then one can see immediately that
b, b′ ∈ X. Additionally, since the component parts of a, b, b′ and x are connected
consequently d(a, b) = d(b, b′ ) = d(b′ , x) = 1. Hence we conclude that d(a, x) 6 3.
Corollary 3.6.2. For n = 2m > 8, DiamC(2. Sn , X) 6 3.
Theorem 3.6.3. Suppose that n > 2m + 2 and n > 8. We have DiamC(2. Sn , X) 6 4.
Proof. For n = 9, 10 and m = 3, we use Magma to calculate then discs structure
of C(2. Sn , X). We find DiamC(2. S9 , X) = 4 with ∆1 (t) = 37, ∆2 (t) = 360, ∆3 (t) =
1602 and ∆4 (t) = 520 where |X| = 2520. Also DiamC(2. S10 , X) = 3 with ∆1 (t) =
91, ∆2 (t) = 1872 and ∆3 (t) = 4336 where |X| = 6300. Thus we may assume that n >
Q1
Q2
Q3
11. For any x ∈ X, it is possible to write x as follows: x = ri=1
[αi αi′ ] ri=1
[δi δi′ ] ri=1
Q4
Q5
[βi βi′ ] ri=1
[γi ] ri=1
[λi ], such that αi , αi′ , βi , γi ∈ supp(a) while βi′ , δi , δi′ , λi ∈ f ix(a).
Also we have r4 + r5 > 2 as n > 2m + 2. For the purpose of finding a path between a
3.6. The Diameter of C(2. Sn , X)
54
and x, we will show there is b ∈ X with no black vertices connected to white vertices of
Gba beside commutes with x. Thus by Proposition 3.6.1 yields the existence of a path
between a and b and hence between a and x. However, to deal with this argument we
will consider the following cases:
i- If r3 = 2s + 1 with s odd, then set
b=(
Qr1
i=1 [αi
αi′ ]
Qr2
i=1 [δi
δi′ ])[β1 β2 ][β3 β4 ] · · · [βr3 −2 βr3 −1 ][β1′ β2′ ][β3′ β4′ ] · · · [βr′ 3 −2
βr′ 3 −1 ][µ1 µ2 ] where µ1 , µ2 ∈ supp(a) ∩ f ix(x). In this case the total number of
,
in Gxb is {s, 1}.
ii- If r3 = 2s + 1 and s even, we have the following cases:
• if r1 = r2 , then we consider to possibilities:
(a) If r3 = 1, then since n > 11, we let b = [α1 α1′ ][µ1 µ2 ][µ3 µ4 ] where
µ1 , µ2 ∈ supp(a) ∩ f ix{x} and µ3 , µ4 ∈ f ix(a) ∩ f ix{x}. Thus the total
,
number of
in Gxb {0, 2}.
(b) if r3 6= 1, we set
b = [α1 α1′ ][β1 β2 ][β3 β4 ] · · · [βr3 −2 βr3 −1 ][β1′ β2′ ]
[β3′ β4′ ] · · · [βr′ 3 −2 βr′ 3 −1 ][µ1 µ2 ][µ3 µ4 ],
where µ1 , µ2 , µ3 , µ4 ∈ supp(a) ∩ f ix(x). Then b commutes with x. The
total number of
,
in Gxb is {s, 2}.
• If r1 + r2 = 1, then x is not an involution element in this case this because
r1 + r2 + r3 = 2s + 2 = m and m not congruent to either 0,1 or 3 mod 4 in
this case contradiction to Proposition 3.1.6.
• If at least one of r1 > 1 or r2 > 1, then we may assume without loss of
generality that r1 > 1. Hence
b=(
Q r1
i=3 [αi
αi′ ]
Q r2
i=1 [δi
δi′ ])[β1 β2 ][β3 β4 ] · · · [βr3 −2 βr3 −1 ]
[β1′ β2′ ][β3′ β4′ ] · · · [βr′ 3 −2 βr′ 3 −1 ][µ1 µ2 ][α1 α2 ][α1′ α2′ ]
where µ1 , µ2 ∈ supp(a) ∩ f ix(x). Therefore, the total number of
in Gxb equals {s + 1, 1}.
,
Chapter 3. Commuting Involution Graph of 2. Sn
55
• If r1 = r2 = 0, then n > 3(2s + 1)(this because supp(x) = 2(2s + 1) and
f ix(x) > 2s + 1) and to find a path between a and x we need to consider
the following:
(a) If n = 3(2s + 1), then we may write x as follows x = [1 2m + 1][3 2m +
2] · · · [2m − 3 n − 1][2m − 1 n], and we set b1 = [1 3][5 7] · · · [2m − 5 2m −
3][2m+1 2m+2] · · · [n−2 n−1][2m−1 n], b2 = [1 3][5 7] · · · [2m−5 2m−
3][2 4][6 8] · · · [2m−4 2m−2][2m−1 n] and b3 = [1 3][2 4][5 6][7 8] · · · [2m−
3 2m − 2][2m + 1 2m + 2]. Thus the total number of
,
in
x , G b1 , G b2 , G b3 receptively {s, 0}, {0, s}, {s − 1, 1}, {1, 1}.
Gb1
a
b2
b3
(b) If n > 3(2s + 1), then set
b = [β1 β2 ][β3 β4 ] · · · [βr3 −4 βr3 −3 ][β1′ β2′ ][β3′ β4′ ] · · · [βr′ 3 −4 βr′ 3 −3 ]
[µ1 µ2 ][µ3 µ4 ][µ5 µ6 ]
where µ1 , µ2 , µ3 , µ4 , µ5 , µ6 ∈ supp(a) ∩ f ix(x), so that the total number
,
of
in Gxb is {s − 1, 3}.
iii- If r3 = 2s with s odd then put
b=(
Qt1
i=1 [αi
αi′ ]
Qt2
i=1 [δi
δi′ ])[β1 β2 ][β3 β4 ] · · · [βr3 −1 βr3 ]
[β1′ β2′ ][β3′ β4′ ] · · · [βr′ 3 −1 βr′ 3 ][µ1 µ2 ]
where µ1 , µ2 ∈ supp(a) ∩ f ix(x). Here t1 = r1 − 1 , t2 = r2 if r1 6= 0 , and if
r1 = 0 then set t2 = r2 − 1. The total number of
,
in Gxb is {s, 1}.
iv- If r3 = 2s with s even then put
b=(
Q r1
i=1 [αi
αi′ ]
Q r2
i=1 [δi
δi′ ])[β1 β2 ][β3 β4 ] · · · [βr3 −1 βr3 ]
[β1′ β2′ ][β3′ β4′ ] · · · [βr′ 3 −1 βr′ 3 ]. Here the total number of
{s, 0}.
,
in Gxb equals
3.6. The Diameter of C(2. Sn , X)
56
Theorem 3.6.4. Suppose that n = 2m + 2 > 10. Then DiamC(2. Sn , X) 6 4.
Proof. Assume that {β1 , β2 } = f ix(a) and let x ∈ X. If f ix(a) = f ix(x) or
x[β1 β2 ] = ±x, so no black vertices are connected with white vertices, by Proposition
3.6.1, we get DiamC(2. Sn , X) 6 4. Additionally, if β1x 6= β2 and {β1 , β2 } ∈ supp(x)
then in a similar way as case (iii) in Theorem 2.6.3, we obtain DiamC(2. Sn , X) 6 4.
Last but not least, if f ix(x) = {β1 , α} or {β2 , α} where α is not equal to neither β1
or β2 , then without loss of generality we assume that f ix(x) = {β2 , α}. To find a path
between a and x, we assume that C1 , C2 , · · · , Ch are the connected components of Gx .
Then all Cj ′ s are cycle or loops except one. We denote this components by Ci . Hence
.
Ci is a chain starting with β1 and ending with black vertices,
Let ai
and xi be the generators of Ci in Gx , so we may assume without loss of
generality that for k > 1, xi = [2 3][4 5][6 7] · · · [2k − 2
2k − 1][1
β1 ][2k] and
ai = [1 2][3 4] · · · [2k − 1 2k][β1 ]. Now let al = [β2 ] and xl = [β2 ]. Also let ail =
[1 2][3 4] · · · [2k − 1 2k][β1 ][β2 ] and xil = [2 3][4 5][6 7] · · · [2k − 2 2k − 1][1 β1 ][2k][β2 ].
For k = 1 we may let ail = [1, 2][β1 ][β2 ] and xil = [1 β1 ][2][β2 ] for k = 1. Since all Cj ′ s
except Ci have no white vertices connected with black vertices, by Proposition 3.6.1 it
suffices to find a path between ail and xil to find a path between a and x, and for
that reason we need to look at the following cases:
Case I: If k = 1, then ail = [1, 2][β1 ][β2 ] and xil = [1 β1 ][2][β2 ], and Gxailil is given by
. Thus in order to find the path between ail and xil , we have to consider
the following cases:
i- If Gx has a connected component Cj such that Cj is a cycle of length 2, we
may assume that aj = [3 4][5 6] and xj = [3 5][4 6] are the corresponding
parts of a and x respectively, that is they generate Cj in Gx . Now put
a
ilj
ailj = [1 2][3 4][5 6] and xilj = [1 β1 ][3 5][4 6], so that Gxilj
is given by
. Now if n = 10 then we my let a = [1 2][3 4][5 6][7 8]
and x = [1 9][3 5][4 6][7 8]. Taking b = [1 2][3 5][4 6][9 10] and b∗ =
[1 9][3 6][4 5][2 10], we have a path from a to x. On the other hand, if
Chapter 3. Commuting Involution Graph of 2. Sn
57
n > 10 it is possible to consider the following subcases:
i. If Gx has connected component Cp such that Cp is a cycle of length 2,
we may assume that ap = [7 8][9 10] and xp = [7 9][8 10] are the corresponding parts that generate Cp in Gx . Set ailjp = [1 2][3 4][5 6][7 8][9 10]
and xiljp = [1 β1 ][3 5][4 6][7 9][8 10]. Let biljp = [1 2][3 5][4 6][7 8][β1 β2 ],
a
b
b
∗
iljp
iljp
iljp
b∗iljp = [1 β1 ][2 β2 ][4 6][7 8][9 10]. Then Gbiljp
,Gbiljp
∗ , Gxiljp are presented
, and hence we get
as
d(ailjp , biljp ) = d(biljp , b∗iljp ) = d(b∗iljp , xiljp ) = 1.
ii. If Gx has connected component Cp such that Cp is a cycle of length
, we may assume that ap = [7 8][9 10][11 12] and
3,
xp =
[8 9][10 11][7 12] are the corresponding parts that generate Cp in Gx .
Set ailjp = [1 2][3 4][5 6][7 8][9 10][11 12] and xiljp = [1 β1 ][3 5][4 6][8 9]
[10 11][7 12]. Let biljp = [1 2][3 4][5 6][8 9][7 10][β1 β2 ],b∗iljp = [1 β1 ][2 β2 ]
[3 4][5 6][7 10][11 12] and b∗∗
iljp = [1 β1 ][2 β2 ][3 5][4 6][7 11][10 12]. Then
a
b
b
iljp
Gbiljp
,Gb∗iljp , Gxiljp
iljp
iljp
b
∗∗
given by
∗
iljp
while Gbiljp
∗∗ representative as follows
.
iii. If Gx has a connected component Cp such that Cp is a cycle of length
greater than 3, let ap = [7 8][9 10] · · · [2r + 5 2r + 6]
and xp =
[8 9][10 11] · · · [2r + 4 2r + 5][7 2r + 6] where r > 3. Now put ailjp =
[1 2][3 4][5 6][7 8] · · · [2r+5 2r+6] and xiljp = [1 β1 ][3 5][4 6][8 9][10 11] · · ·
a
iljp
is given by
[2r + 4 2r + 5][7 2r + 6]. The graph Gxiljp
, so that our aim is to find
biljp , b∗iljp and b∗∗
iljp (if required) which connect ailjp to xiljp . For that
reason we argue the following cases:
• If r is even such that r = 2s and s is even, then set biljp = [1 2][3 4][5 6]
[7 2r+6][8 2r+5] · · · [r+5 r+8][r+6 r+7], b∗iljp = [1 2][3 6][4 5][7 2r+
6][8 2r+5] · · · [r+5 r+8][β1 β2 ] and b∗∗
iljp = [1 β1 ][2 β2 ][3 5][4 6][7 2r+
6][8 2r + 5] · · · [r + 5 r + 8]. Thus the total number of
,
in
3.6. The Diameter of C(2. Sn , X)
a
b
58
b
∗
b
iljp
iljp
iljp
and Gxiljp
Gbiljp
, Gbiljp
∗ , Gb∗∗
iljp
∗∗
respectively {s, 0},{1, 1},{2, 0} and
iljp
{s − 1, 1}.
• If r is even such that r = 2s
and
s is odd, then set biljp =
[1 2][3 5][4 6][7 2r + 6][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7], b∗iljp =
[1 2][3 6][4 5][7 2r + 6][8 2r + 5] · · · [r + 5 r + 8][β1 β2 ] and b∗∗
iljp =
[1 β1 ][2 β2 ][3 4][5 6][7 2r + 6][8 2r + 5] · · · [r + 5 r + 8]. The total
number of
a
b
b
∗
b
iljp
iljp
iljp
, in Gbiljp
and Gxiljp
, Gbiljp
∗ , Gb∗∗
iljp
∗∗
iljp
respectively
{s + 1, 0},{1, 1},{2, 0} and {s, 1}.
• If r is odd such that r = 2s + 1 and s is odd, then set biljp =
[1 2][3 5][4 6][7 2r + 6][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7], b∗iljp =
[1 2][3 6][4 5][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7][β1 β2 ] and
b∗∗
iljp = [1 β1 ][2 β2 ][3 5][4 6][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7]. Here
,
the total number of
a
∗
b
b
b
iljp
iljp
iljp
and Gxiljp
in Gbiljp
, Gbiljp
∗ , Gb∗∗
iljp
∗∗
iljp
respectively {s + 1, 0},{1, 1},{2, 0} and {s, 1}.
• If r is odd such that r = 2s + 1 and s is even,then set biljp =
[1 2][3 4][5 6][7 2r + 6][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7], b∗iljp =
[1 2][3 6][4 5][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7][β1 β2 ] and
b∗∗
iljp = [1 β1 ][2 β2 ][3 4][5 6][8 2r + 5] · · · [r + 5 r + 8][r + 6 r + 7]. Then
,
the total number of
a
b
∗
b
b
iljp
iljp
iljp
in Gbiljp
, Gbiljp
and Gxiljp
∗ , Gb∗∗
iljp
∗∗
iljp
respectively {s, 0},{1, 1},{2, 0} and {s + 1, 1}.
ii- If Gx has a connected component Cj such that Cj is a cycle of length
3,
, let aj = [3 4][5 6][7 8]
put ailj = [1 2][3 4][5 6][7 8] and
and
xj = [4 5][6 7][3 8].
Here
xilj = [1 β1 ][4 5][6 7][3 8]. In this
case let bilj = [β1 β2 ][3 8][5 6][4 7], b∗ilj = [1 2][β1 β2 ][4 5][6 7] and
b∗∗
ilj = [2 β2 ][1 β1 ][4 7][5 6].
a
b
b
ilj
ilj
ilj
Then Gbilj
, Gbilj
∗ and Gxilj
b
∗∗
are given by
∗
ilj
. On the other hand Gbilj
∗∗ is presented by
. Hence by using Lemma 3.4.2 we obtain d(ailj , bilj ) =
∗∗
d(bilj , b∗ilj ) = d(b∗ilj , b∗∗
ilj ) = d(bilj , xilj ) = 1.
Chapter 3. Commuting Involution Graph of 2. Sn
59
iii- If Gx has a connected component Cj such that Cj is a cycle of length more
than 3, let aj = [3 4][5 6] · · · [2r + 1 2r + 2] and xj = [4 5][6 7] · · · [2r 2r +
1][3 2r + 2]. Now put ailj = [1 2][3 4][5 6][7 8] · · · [2r + 1 2r + 2] and
a
ilj
xilj = [1 β1 ][4 5][6 7] · · · [2r 2r + 1][3 2r + 2]. The graph Gxilj
is given by
, so that our purpose is to find bilj , b∗ilj and
b∗∗
ilj (if required) which link ailj and xilj . For that intent we look at the
following cases:
• If r is even such that r = 2s and s is even, then set bilj = [1 2][3 2r +
2][4 2r + 1][5 2r] · · · [r + 1 r + 4][r + 2 r + 3], b∗ilj = [1 2][β1 β1 ][4 5][2r +
1 2r][6 2r − 1] · · · [r + 2 r + 3] and b∗∗
ilj = [1 β1 ][2 β2 ][4 2r + 1][5 2r] · · · [r +
a
and
b ∗∗
Gxilj
ilj
b
b
ilj
ilj
ilj
, in Gbilj
, Gbilj
∗ , Gb∗∗
1 r+4][r+2 r+3]. Thus the total number of
∗
ilj
respectively equals {s, 0},{1, 1},{2, 0} and {s − 1, 1}.
• If r is even such that r = 2s and s is odd, then set bilj = [β1 β2 ][3 2r +
2][4 2r + 1][5 2r] · · · [r + 1 r + 4][r + 2 r + 3], b∗ilj = [1 2][β1 β1 ][4 2r][2r +
1 5][6 2r − 1] · · · [r + 2 r + 3] and b∗∗
ilj = [1 β1 ][2 β2 ][4 5][2r 2r + 1][6 2r −
1] · · · [r + 2 r + 3]. The total number of
and
b ∗∗
Gxilj
ilj
,
a
b
b
ilj
ilj
ilj
in Gbilj
, Gbilj
∗ , Gb∗∗
∗
ilj
respectively {s, 1},{1, 1},{2, 0} and {s − 2, 1}.
• If r is odd such that r = 2s+1 and s is odd,then set bilj = [β1 β2 ][3 2r +
2][4 2r + 1][5 2r] · · · [r + 1 r + 4][r + 2 r + 3], b∗ilj = [1 2][β1 β1 ][4 5][2r +
1 2r][6 2r − 1] · · · [r + 2 r + 3] and b∗∗
ilj = [1 β1 ][2 β2 ][4 2r + 1][5 2r] · · · [r +
1 r + 4][r + 2 r + 3]. Here the total number of
b
Gxilj
ilj
∗∗
b
b
∗
,
a
ilj
in Gbilj
, and
ilj
ilj
is {s, 1}, while in Gbilj
respectively equals {1, 1},{2, 0}.
∗ , Gb∗∗
ilj
• If r is odd such that r = 2s + 1 and s is even, then set bilj = [1 2][3 2r +
2][4 2r + 1][5 2r] · · · [r + 1 r + 4][r + 2 r + 3], b∗ilj = [1 2][β1 β1 ][4 2r][2r +
1 5][6 2r − 1] · · · [r + 2 r + 3] and b∗∗
ilj = [1 β1 ][2 β2 ][4 5][2r 2r + 1][6 2r −
1] · · · [r + 1 r + 4][r + 2 r + 3]. Thus the total number of
a
b
b
∗
b
ilj
ilj
ilj
and Gxilj
in Gbilj
, Gbilj
∗ , Gb∗∗
ilj
ilj
1, 1}.
∗∗
,
respectively {s, 0},{1, 1},{2, 0} and {s −
3.6. The Diameter of C(2. Sn , X)
60
iv- If ω1 and ω2 be the number of cycles of length 2 and 3 and Cω1 and Cω2
be their connected component in Gx respectively. We need to consider the
following sub cases:
• If (ω1 , ω2 ) = (3, 0) then we suppose without loss of generality that, aω1 =
[1 2][3 4][5 6][7 8][9 10][11 12][13 14] and xω1 = [1 β1 ][4 5][3 6][8 9][7 10]
[12 13][11 14]. Hence if we set bω1 = [1 2][4 5][3 6][7 8][9 10][12 13][11 14],
b∗ω1 = [1 2][β1 β2 ][3 6][7 8][9 10][11 12][13 14] and b∗∗
ω1 = [1 β1 ][2 β2 ][3 6]
[7 8][9 10][12 13][11 14], then d(aω1 , bω1 ) = d(bω1 , b∗ω1 ) = d(b∗ω1 , b∗∗
ω1 ) =
d(b∗∗
ω1 , xω1 ) = 1.
aω
bω
bω
bω
Gbω 1 , Gb∗∗1 equals {2, 0}, and in Gbω 1 ∗ , Gxω11
1
,
This is because the total number of
∗
ω1
1
∗∗
in
equals {1, 1}. Moreover, If
(ω1 , ω2 ) equal to the one of the followings (1, 0), (0, 1), (2, 1) and (1, 2).
Then by previous cases one can see immediately that there is a edges
links the components of x and a. In general, if ω1 + ω2 = s and s is odd
then there is always a path between a and x. This because s = 3 + r1
where r1 is an even, thus by the former cases there exist a path between
the components of x and a.
• If (ω1 , ω2 ) = (0, 2) then we suppose without loss of generality that, aω2 =
[1 2][3 4][5 6][7 8][9 10][11 12][13 14] and xω2 = [1 β1 ][4 5][6 7][3 8][10 11]
[12 13][9 14]. Hence if we set bω2 = [1 2][3 8][4 7][5 6][9 14][10 13][11 12],
b∗ω2 = [1 2][β1 β2 ][4 6][5 7][9 14][10 13][11 12] and b∗∗
ω2 = [1 β1 ][2 β2 ][4 5]
[6 7][9 14][10 13][11 12]. The total number of
equals {2, 0}, and in
bω
bω ∗∗
Gb∗ 2 , Gxω22
ω2
,
aω
bω
in Gbω 2 , Gb∗∗2
2
∗
ω2
equals {1, 1}. We get d(aω2 , bω2 ) =
∗∗
d(bω2 , b∗ω2 ) = d(b∗ω2 , b∗∗
ω2 ) = d(bω2 , xω2 ) = 1. Additionally, if (ω1 , ω2 )
equal (2, 0) or (1, 1) then by the same argument as above we can find a
path in this case. Overall, if ω1 + ω2 = s and s is even , set s = 2 + r1
where r1 in this case even. Thus the existence of the path is satisfied.
Case II: If k = 2, then without loss of generality, we may set ail = [1 2][3 4][β1 ][β2 ], and
xil = [2 3][1 β1 ][4][β2 ]. In order to find the path between ail and xil , we have to
Chapter 3. Commuting Involution Graph of 2. Sn
61
consider the following cases:
i- If Gx has a connected component Cj such that Cj is a cycle of length 2,
we may assume that aj = [5 6][7 8] and xj = [6 7][8 5] are the corresponding parts of a and x respectively, that is they generate Cj in Gx .
Now put ailj = [1 2][3 4][5 6][7 8] and xilj = [2 3][1 β1 ][6 7][8 5], and
let bilj = [1 4][2 3][5 8][6 7], b∗ilj = [1 4][β1 β2 ][5 6][7 8]
a
and
∗
b
ilj
ilj
[1 β1 ][4 β2 ][5 7][6 8]. Thus Gbilj
, Gbilj
∗∗ are given by
b
b
iljp
iljp
whilst Gbiljp
∗ , Gxiljp
∗∗
b∗∗
ilj =
,
. So we obtain
are represented as
∗∗
d(ailj , bilj ) = d(bilj , b∗ilj ) = d(b∗ilj , b∗∗
ilj ) = d(bilj , xilj ) = 1.
ii- If Gx has a connected component Cj such that Cj is a cycle of length 3,
, then let aj = [5 6][7 8][9 10] and xj = [6 7][8 9][10 5]. Here put
ailj = [1 2][3 4][5 6][7 8][9 10] and xilj = [2 3][1 β1 ][6 7][8 9][10 5]. In this
case let bilj = [1 4][2 3][5 6][7 10][8 9] , b∗ilj = [1 4][β1 β2 ][5 6][7 9][8 10] and
a
b
∗
ilj
ilj
b∗∗
ilj = [4 β2 ][1 β1 ][5 6][8 9][7 10]. Therefore, Gbilj , Gbilj ∗∗ are given by
b
b
, while Gb∗ilj , Gxilj
ilj
ilj
∗∗
are represented by
. This yields d(ailj , bilj ) = d(bilj , b∗ilj ) = d(b∗ilj , b∗∗
ilj )
= d(b∗∗
ilj , xilj ) = 1.
iii- If Gx has a connected component Cj such that Cj is a cycle of length greater
than 3, then without loss of generality let aj = [5 6][7 8] · · · [2r + 3 2r +
4]
and
xj = [6 7][8 9] · · · [5 2r + 4] where r > 3.
Also let ailj =
[1 2][3 4][5 6][7 8] · · · [2r + 3 2r + 4] and xilj = [2 3][1 β1 ][6 7][8 9] · · · [2r +
a
ilj
2 2r + 3][5 2r + 4]. Then the graph Gxilj
is given by
, so that our purpose is to find bilj , b∗ilj
and b∗∗
ilj which form a path between ailj and xilj . For that we check the
following cases:
• If r is even such that r = 2s and s is even, then set bilj = [2 3][1 4][5 6][7 2r
+4][8 2r + 3] · · · [r + 4 r + 7][r + 5 r + 6], b∗ilj = [1 4][β1 β2 ][5 7][6 2r +
4][8 2r + 3] · · · [r + 4 r + 7][r + 5 r + 6] and b∗∗
ilj = [4 β2 ][1 β1 ][6 7][5 2r +
3.6. The Diameter of C(2. Sn , X)
62
4][8 2r + 3] · · · [r + 4 r + 7][r + 5 r + 6]. Then the total number of
a
b
b
∗
b
ilj
ilj
ilj
and Gxilj
, in Gbilj
, Gbilj
∗ , Gb∗∗
ilj
∗∗
ilj
respectively {s, 0},{1, 1},{2, 0}
and {s − 1, 1}.
• If r is even such that r = 2s and s is odd, then let bilj = [2 3][1 4][5 r +
5][6 r + 6][7 2r + 4][8 2r + 3] · · · [r + 3 r + 8][r + 4 r + 7], b∗ilj =
[1 4][β1 β2 ][5 r + 6][6 r + 5][7 2r + 4][8 2r + 3] · · · [r + 3 r + 8][r + 4 r + 7]
and bilj ∗∗ = [4 β2 ][1 β1 ][5 6][7 2r +4][8 2r +3] · · · [r +4 r +7][r +5 r +6].
,
Thus the total number of
a
b
b
∗
b
ilj
ilj
ilj
in Gbilj
, Gbilj
and Gxilj
∗ , Gb∗∗
ilj
∗∗
ilj
re-
spectively {s + 1, 0},{1, 1},{2, 0} and {s, 1}.
• If r is odd such that r = 2s + 1
and
s is odd, then let bilj =
[2 3][1 4][5 6][7 2r+4][8 2r+3] · · · [r+3 r+8][r+4 r+7][r+5 r+6], b∗ilj =
[1 4][β1 β2 ][5 7][6 2r + 4][8 2r + 3] · · · [r + 3 r + 8][r + 4 r + 7][r + 5 r + 6]
and b∗∗
ilj = [4 β2 ][1 β1][5 6][7 2r + 4][8 2r + 3] · · · [r + 4 r + 7][r + 5 r + 6].
a
b
∗
b
b
ilj
ilj
ilj
and Gxilj
, in Gbilj
, Gbilj
∗ , Gb∗∗
ilj
Therefore, the total number of
∗∗
ilj
respectively {s + 1, 0},{1, 1},{2, 0} and {s, 1}.
• If r is odd such that r = 2s + 1
and
s is even, then let bilj =
[1 2][3 4][5 6][7 2r + 4][8 2r + 3] · · · [r + 4 r + 7][r + 5 r + 6], b∗ilj =
[1 2][3 4][5 6][7 2r + 3][8 2r + 4][9 2r + 2] · · · [r + 4 r + 7][β1 β2 ] and
b∗∗
ilj = [1 2][3 β1 ][4 β2 ][5 6][7 2r + 4][8 2r + 3][9 2r + 2] · · · [r + 4 r + 7].
The total number of
,
a
b
b
∗
b
ilj
ilj
ilj
in Gbilj
and Gxilj
, Gbilj
∗ , Gb∗∗
ilj
ilj
∗∗
respec-
tively {s, 0},{1, 1},{2, 0} and {s + 1, 1}.
iv- If ω1 and ω2 are the number of cycles of length 2 and 3 respectively, then
we need to consider the following sub cases:
• If ω1 = ω2 = 1 then we suppose without loss of generality that Cω1
and Cω2 are the corresponding connected components in Gx . Let aω1 =
[5 6][7 8] and aω2 = [9 10][11 12][13 14]. Also let xω1 = [6 7][5 8] and
xω2 = [10 11][12 13][9 14]. Hence set ailω1 ω2 = [1 2][3 4][5 6][7 8][9 10][11
12][13 14][β1 ][β2 ] and xilω1 ω2 = [2 3][1 β1 ][6 7][5 8][10 11][12 13][9 14][4]
[β2 ], we may set ξ = ilω1 ω2 . Set bξ = [2 3][1 4][5 6][7 8][9 10][12 13][11 14],
Chapter 3. Commuting Involution Graph of 2. Sn
63
b∗ξ = [β1 β2 ][1 4][5 7][6 8][9 10][12 13][11 14] and b∗∗
ξ = [1 β1 ][4 β2 ][6 7][5 8]
[9 10][12 13][11 14]. Since the total number of
equals {2, 0}, and in
b
b ∗∗
Gb∗ξ , Gxξξ
ξ
a
b
∗
ξ
in Gbξξ , Gb∗∗
,
ξ
equals {1, 1}, we obtain that d(aξ , bξ ) =
∗∗
d(bξ , b∗ξ ) = d(b∗ξ , b∗∗
ξ ) = d(bξ , xξ ) = 1.
• If ω1 = 0 and ω2 = 2 or vice versa, we may assume without loss of
generality that ω1 = 2 and ω2 = 0 (if ω1 = 0 and ω2 = 2 then by the
same way a path between x and b can be found). Then suppose that
Cω1 and Cω1 ∗ are the connected components in Gx , with component
parts aω1 = [5 6][7 8] , aω1∗ = [9 10][11 12], xω1 = [6 7][5 8] and xω1∗ =
[10 11][9 12]. Thus if we let ailω1 ω1∗ = [1 2][3 4][5 6][7 8][9 10][11 12] and
xilω1 ω1∗ = [2 3][1 β1 ][6 7][5 8][10 11][9 12], we may let ξ = ilω1 ω1∗ . Put
bξ = [2 3][1 4][5 7][6 8][9 10][11 12], b∗ξ = [β1 β2 ][1 4][5 7][6 8][10 11][9 12]
and b∗∗
ξ = [1 β1 ][4 β2 ][5 6][7 8][10 11][9 12]. We get that d(aξ , bξ ) =
d(bξ , b∗ξ ) = d(b∗ξ , b∗ξ ) = d(b∗∗
ξ , xξ ) = 1, as one can see that the total
number of
,
a
b
∗
b
b
ξ
in Gbξξ , Gb∗∗
equals {2, 0}, and in Gb∗ξ , Gxξξ
ξ
ξ
∗∗
equals
{1, 1}.
• If ω1 + ω2 = s and s odd, then there is a path in this case. This because
if s = 1 then by previous cases the path exists and since s = r + 1 where
r is even and again the path exists by the Case 4.1. Whilst if s is even,
then if s = 2 by the above there is a path between a and x. Moreover,
if s > 2, then s = 2 + q, such that q is an even and again by the above
yields the existence of a path between the components of a and b such
that d(a, x) 6 4.
Case III: If k > 2, then Gxailil is given by
. Our aim is to find bil , b∗il
and b∗∗
il (if required) which build a path between ail and xil . For that purpose we
need to take into account the following cases:
i- If k = 2s and s is even, then if s = 2 we let bil = [β1 β2 ][3 8][4 7][5 6],
b∗il = [3 β1 ][8 β2 ][4 5][6 7] and b∗∗
il = [3 β1 ][12][4 7][5 6] . Otherwise take
3.6. The Diameter of C(2. Sn , X)
64
bil = [β1 β2 ][3 2k][4 2k − 1][5 2k − 2] · · · [k − 1 k + 4][k k + 3][k + 1 k + 2] ,
b∗il = [2k β2 ][3 β1 ][4 2k − 1][5 2k − 2] · · · [k − 1 k + 4][k k + 1][k + 2 k + 3]
and b∗∗
il = [1 2][3 β1 ][4 2k − 1][5 2k − 2] · · · [k − 1 k + 4][k k + 3][k + 1 k + 2].
∗
il
and Gxbilil
, in Gbailil , Gbbilil∗ , Gbb∗∗
Thus the total number of
∗∗
il
respectively
{s − 1, 1},{2, 0}, {1, 1} and {s, 0}.
ii- If k = 2s and s is odd, then if s = 3 set bil = [β1 β2 ][3 12][4 11] [5 6][9 10][7 8],
b∗il = [12 β2 ][3 β1 ][4 11][5 10][6 9][7 8] and b∗∗
il = [1 2][3 β1 ][4 11][5 10][6 7]
[8 9] . Otherwise take bil = [β1 β2 ][3 2k][4 2k − 1][5 2k − 2] · · · [k − 2 k +
5][k − 1 k][k + 3 k + 4][k + 1 k + 2], b∗il = [2k β2 ][3 β1 ][4 2k − 1][5 2k −
2][6 2k − 3] · · · [k − 2 k + 5][k − 1 k + 4][k k + 3][k + 1 k + 2] and b∗∗
il =
[1 2][3 β1 ][4 2k −1][5 2k −2] · · · [k −1 k +4][k k +1][k +2 k +3]. One can show
∗
il
, in Gbailil , Gbbilil∗ , Gbb∗∗
and Gxbilil
that the total number of
il
∗∗
respectively
{s − 2, 1},{2, 0}, {1, 1} and {s − 1, 0}.
iii- If k = 2s + 1 and s is odd, then if s = 1 let bil = [β1 β2 ][3 6][4 5] and
b∗il = [1 2][3 β1 ][6 β2 ] . Otherwise take bil = [k + 1 k + 2][3 2k][4 2k −
1][5 2k − 2] · · · [k k + 3][β1 β2 ] and b∗il = [1 2][3 β1 ][4 2k − 1][5 2k − 2][6 2k −
3] · · · [k k+3][2k β2 ] . So that that the total number of
, in Gbailil , Gxbilil
∗
equals {s, 1}, while in Gbb∗il it equals {1, 1}.
il
iv- If k = 2s + 1 and s is even, then if s = 2 let bil = [β1 β2 ][3 10][4 9] [5 6][7 8],
b∗il = [3 β1 ][10 β2 ][4 9][5 7][6 8] and b∗∗
il = [1 2][3 β1 ][4 9] [5 8][6 7] .
Otherwise take bil = [β1 β2 ][3 2k][4 2k − 1][5 2k − 2] · · · [k − 1 k + 4][k k +
1][k + 2 k + 3], b∗il = [2k β2 ][3 β1 ][4 2k − 1][5 2k − 2][6 2k − 3] · · · [k − 1 k +
4][k k +2][k +1 k +3] and b∗∗
il = [1 2][3 β1 ][4 2k −1][5 2k −2][6 2k −3] · · · [k −
1 k + 4][k k + 3][k + 1 k + 2]. We can show that the total number of
∗
il
and Gxbilil
in Gbailil , Gbbilil∗ , Gbb∗∗
il
∗∗
,
respectively equals {s − 1, 1},{2, 0}, {1, 1} and
{s, 0}.
Thus we conclude from the above cases and Proposition 3.6.1, that d(a, x) 6 4.
Chapter 3. Commuting Involution Graph of 2. Sn
65
Nevertheless, Theorem 3.6.3 is not satisfied when n = 8 and m = 3. Using Magma
the analysis of the discs structure of t is as follows:
1
t
10
10
13
∆2 (t)
∆1 (t)
1
1
2
8
∆3 (t)
3
2
2
∆4 (t)
5
2
∆5 (t)
3
6
10
Where |∆1 (t)| = 13, |∆2 (t)| = 60, |∆3 (t)| = 294, |∆3 (t)| = 424 and |∆5 (t) = 48.
Furthermore, for xi ∈ ∆i (t), i = 1, 2, 3, 4, 5, we calculate the following: The number
of elements in ∆1 (t) which commute with t, x1 , x2 respectively equals 13, 2, 1. The
number of elements in ∆2 (t) which commute with x1 , x2 , x3 is 10, 2, 3 respectively.
The number of elements in ∆3 (t) which commute with x2 , x3 , x4 is 10, 2, 5 respectively.
Whilst the number of elements in ∆4 (t) which commute with x3 , x4 , x5 respectively
equals 8, 6, 3. Finally, the number of elements in ∆5 (t) which commute with x4 , x5
respectively equals 2, 10.
As we can see from the above calculation the DiamC(2. S8 , X) = 5, where t =
.
[1 2][3 4][5 6] and X = t2 S8 . However, Theorem 3.6.3 shows that DiamC(2. Sn , X) 6 4
for n = 2m + 2 > 10.
Chapter 4
Commuting Involution Graphs of
Double Covers of Sporadic
Groups and Their Automorphism
Groups
4.1
Introduction
Suppose that G is a sporadic simple group and X a conjugacy class of involutions in
G. The commuting involution graphs C(G, X) and C(G.2, X) have been studied by
P.Rowley and his PhD student (Chris Bates, David Bundy and Sarah Hart) in [16]. In
this chapter we consider the commuting involution graphs of double covers of sporadic
simple groups C(2. G, X), and double covers of their automorphism groups C(2. G.2, X).
One can see from the classification of finite simple groups that sporadic simple groups
are as follows:
• M11 , M12 , M22 , M23 , M24 (the Mathieu groups).
• Co1 , Co2 , Co3 , M cL, HS, Suz, J2 (the Leech lattice groups).
• F i22 , F i23 , F i′24 (the Fischer groups).
66
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
67
• M, B, T h, HN, He (the Monstrous groups).
• J1 , J3 , J4 , O′ N, Ru, Ly (the pariahs).
In most of the results that we deal with in this chapter, computational technical methods are used to investigate the commuting involution graphs. Magma [23] and GAP
[41] are used to analyze the disc structure of the graphs. Now, for a fixed involution
t in G, the disc △i (t) is a union of CG (t)-orbits of tG (as CG (t) acts by conjugation
on tG ). Therefore, our aim is to determine these orbits and check which of these lie in
△i (t).
4.2
General Results
Our strategy to obtain a complete set of representatives for the CG (t)-orbits on X is
to consider an arbitrary group element g ∈ G and use a variety of computational algorithms and group theory results to obtain the complete set of representatives. Suppose
x ∈ X. Then the size of the orbit xCG (t) is equal to [CG (t) : CCG (t) (x)]. However,
the computational code “IsConjugate” will be employed in many cases to determine
whether an element y ∈ X belongs to xCG (t) . The following lemma is essential in
determining the discs of the commuting involution graph C(G, X).
Lemma 4.2.1. For a finite group G, each disc ∆i (t) of the commuting involution
graph C(G, X) is a union of particular CG (t)-orbits of X, (where CG (t) acts on X by
conjugation and t ∈ X).
Proof. It is enough to prove that for x ∈ ∆i (t) and z ∈ CG (t), xz ∈ ∆i (t). As
d(t, x) = i, the shortest path between t, x is of length i. Moreover, if we conjugate
the element of this path by z, since z ∈ CG (t) we get another path but this time
between t, xz of length i. If d(t, xz ) < i then conjugating by z −1 we obtain d(t, x) < i,
a contradiction. Therefore xz ∈ ∆i (t).
The next theorem (mentioned in [16]) gives basic results which we use as an important tool to determine the elements of ∆i (t).
4.2. General Results
68
Theorem 4.2.2. Let G be a finite group, t ∈ G an involution and let X = tG the
conjugacy class of t. suppose that x ∈ X and put z = tx, and let m equal the order of
z. The following holds
1. CCG (t) (x) = CG (t) ∩ CG (x) = CCG (z) (t) = CCG (z) (x);
2. m = 2 if and only if x ∈ ∆1 (t);
3. x ∈ ∆2 (t) if m is even, m > 4 and z m/2 ∈ X;
4. d(x, t) > 3 if CCG (z) (x) ∩ X = ∅. Specially, if the order of CCG (z) (x) is odd then
d(x, t) > 3;
5. If m is odd and there are no elements g ∈ G with the property that the order of
g is 2m, g 2 = z and g m ∈ X, then d(x, t) > 3.
Proof. Full proof can be found in [16].
Definition 4.2.3. [55] Let G is a group act on a finite set Ω, for g ∈ G define χ(g) =
|{α ∈ Ω|αg = α}|. The non-negative integer value function χ is called the Permutation
Character associated with the action.
The number of CG (t)-orbits under the action of conjugation on tG can be calculated
by using the character table of the group, as we can seen in the following result:
Proposition 4.2.4. Let G be a group acting transitively on a finite set Ω, with a
permutation character χ. Suppose that α ∈ Ω and that Gα has exactly k orbits on Ω.
Then hχ, χi = k.
Proof. See [55, Corollary 5.15].
The quantity k in Proposition 4.2.3 is called the permutation rank of Gα on Ω.
Therefore, the permutation rank of CG (t) on X is equal to the number of CG (t)-orbits
under the conjugation action on X.
Assume that C is a G-conjugacy class. It is obvious that the set XC = {x ∈ X :
tx ∈ C} under the conjugation action of CG (t) breaks up into suborbits. Hence to
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
69
determine all the suborbits of X, we have to identify the CG (t)-orbits of XC , for all
those C such that XC 6= φ. The next definition gives us the size of the set XC , and
therefore, the size of suborbits of XC :
Definition 4.2.5. Let Ci , Cj and Ck be conjugacy classes of a finite group G. Then
for a fixed element g ∈ Ck , define the set
aijk = |{(gi , gj ) ∈ Ci × Cj | gi gj = g}|.
Then for all possible i, j, k the value aijk is called a class class structure constants
for G.
The next lemma will be used to compute the structure constant for G.
Lemma 4.2.6. [44] Let G be a finite group with n conjugacy classes C1 , C2 , . . . Cn .
Then for all i, j, k we have
aijk =
P
χ(gi )χ(gj )χ(gk )
|G|
χ∈Irr(G)
|CG (gi )||CG (gj )|
χ(1)
where gi , gj and gj are respectively in Ci , Cj and Ck .
Proof. Full proof can be found in [44, Lemma 2.12].
Now, since |XC | = |{(c, x) ∈ C × X | cx = t}|, by employing Lemma 4.2.5 we get
|XC | =
P
χ(g)χ(t)2
|G|
|CG (g)||CG (t)| χ∈Irr(G) χ(1)
Therefore, from the complex character table of G, which is available in GAP character
table library, and using the GAP function “Class Multiplication Coefficient” we
immediately obtain the size of XC . Indeed, at the end of this thesis we provide Magma
code for finding |XC |. The following lemma illustrates how we can employ the known
representative of a CG (t)-orbit to get a new one.
Lemma 4.2.7. [65] Let x ∈ XC , so that z = tx ∈ C and suppose n is a divisor of
the order of z. Let D be the conjugacy class containing z n . We define the element
x(n) = tz n and note the following properties:
4.2. General Results
70
i- x(n) ∈ XD ;
ii- if x, y ∈ X are CG (t)-conjugate then so are x(n) and y (n) ; and
iii- all CG (t)-conjugates of x(n) arise as y (n) for some CG (t)-conjugate y of x.
Proof. Full proof can be seen in [65].
Let G be a finite group whose a centre is a cyclic group Z = hzi of order 2. If
θ : G −→ G/Z is the natural homomorphism then we see from Theorem 3.3.1 that for
any conjugacy class C̄ of Ḡ = G/Z, the inverse image θ−1 (C̄) is either a conjugacy
class of G or a union of two classes in G. Moreover, the second case appears exactly
when there is no element of θ−1 (C̄) conjugate to z times itself. Now, under the above
conditions one can show the following:
Lemma 4.2.8. Let C̄ be a conjugacy class of involutions in Ḡ = G/Z, and suppose
that θ−1 (C̄) = C1 ∪ C2 where C1 , C2 are conjugacy class of involutions in G. Then
C(G, C) ∼
= C(Ḡ, C¯1 ) ∼
= C(Ḡ, C¯2 ).
Proof. We only give the proof for the class C¯1 . The class C¯2 is proved in a similar
way. For a ∈ C¯1 we only need to show that x ∈ ∆1 (a) if and only if x̄ ∈ ∆1 (ā) where
ā and x̄ are the lifting of a, x in C¯1 respectively. The covering map implies that if
d(a, x) = 1 then d(ā, x̄) = 1. Conversely assume that d(ā, x̄) = 1. If d(a, x) 6= 1, we
have [a, x] = z, but this mean that a is conjugate to za which contradicts Theorem
3.3.1. Hence we must have d(a, x) = 1. Thus we get our result.
Our main target is to determine the commuting involution graph for all involution
classes X̄ of sporadic groups and their automorphism groups, with the condition that
the inverse image of X̄ under the covering map (which we call it X) is equal to the
union of involution classes in the double cover of the sporadic group or their automorphism groups. Otherwise we will end up with same situation as in Lemma 4.2.7.
The following table provides us the complete set of conjugacy classes of involutions,
which satisfy the above conditions with their permutation ranks.
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
71
Table 4.1: Disc sizes and Permutation Character
Group
2.M12 .2
2.M22 .2
2.Suz.2
2.Hs.2
2.Hs.2
2.Co1
2.F i22
2.F i22 .2
2.F i22 .2
2.F i22 .2
Class
2D
2F
2D
2B
2E
2D
2F
2B
2C
2F
Size of Class
1584
2772
4717440
2200
46200
21361158000
72972900
7020
123552
38918880
Permutation Rank
27
28
115
8
63
465
438
6
7
106
For the remainder of this chapter we give information regarding the computational
techniques involved in each case. We also give tables of the sizes of CG (t)-orbits inside
the conjugacy classes of G, which means the size of sets ∆i (t) ∩ XC , where C is a
G-conjugacy class. We should also refer the reader to [71] for general group-theoretic
definitions and results, related to these groups.
4.3
The Mathieu Groups
In the 19th century Emile Mathieu discovered the Mathieu groups which are the first
family of sporadic simple groups (see [50, 51]). The largest Mathieu group, M24 , may
be defined as an automorphism group of the Steiner system S(5, 8, 24), while the groups
M23 and M22 can be defined to be the point and pointwise stabilizer subgroups in M24
respectively. The groupM12 is a subgroup of M24 and stabilizes a dodecad(a 12-elements
subset) of S(5, 8, 24). Finally, the group M11 is defined as a point Stabilizer subgroup
of M12 . The mathematical tool that may be used to deal with M24 is the Miracle
Octad Generator. For more information about the MOG see [33]. According to the
Table 4.1 we are interested in the cases of (2. M12 .2 , 2D) and (2. M22 .2 , 2F ). Now we
outline the procedure of finding the ∆i (t) ∩ XC for those groups. To do so we create the
following algorithm to finding suborbit representatives and decide which one of these
4.3. The Mathieu Groups
72
representatives lie in ∆i (t). The structure of this algorithm depends on the ability to
calculate CG (t)-orbits and it is summarized as follows :
Algorithm 2
.
Input: G is either 2.M 12.2 or 2.M 22.2, t an involution in 2D or 2F respectively;
i: r −→ Random(tG {t}), set Reps −→ {r}
ii: f or x ∈ tG {t} check if
and
CoR −→ rCG (t) .
x ∈
/ CoR, then CoR 7−→ CoR ∪ {xCG (t) }; and
Reps 7−→ Reps ∪ {x}.
iii: for y in Reps do: if y commutes with t then ∆1 (t) 7−→ ∆1 (t) ∪ {y CG (t) }.
iv: for y in Reps{∆i−1 (t) ∪ ∆i−2 (t) ∪ · · · ∆i−(i−1) (t)}, where ∆0 (t) = {t} do: if
there exist y1 ∈ ∆i−1 (t) ∩ CG (y) then ∆i (t) 7−→ ∆i (t) ∪ {y CG (t) }.
v: if |{∆1 (t) ∪ ∆2 (t) ∪ · · · ∆(j) (t)}| =
|G|
|CG (t)| ,
then the diameter of the graph is
equal to j, where j ∈ N.
vi: for x ∈ ∆i (t) do: if tx ∈ C
then x ∈ XC for G-conjugacy class C.
Output: the set of suborbit representatives and ∆i (t) structure.
Now we employ Algorithm 2 to get the structure of C(G, X), when G is either
2.M 12.2 or 2.M 22.2 and t is an involution in 2D or 2F respectively. In the following
tables we give the size of suborbits of CG (t)-orbits on X = tG inside the classes of G
intersection ∆i (t):
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
73
Table 4.2: C(2. M12 .2, 2D)
Class
2A
2BC
3A
3B
4A
5B
6A
6B
6CD
10A
11A
12A
20A
22A
∆1 (t)
1
15
−
−
−
−
−
−
−
−
−
−
−
−
∆2 (t)
−
−
−
60
−
60, 60
−
60
−
60, 60
−
−
−
−
∆3 (t)
−
−
−
−
−
−
−
−
60, 60
−
120, 120
−
120, 120
120, 120
∆4 (t)
−
−
20, 20
−
30
−
20, 20
−
−
−
−
120
−
−
∆5 (t)
−
−
−
−
2
−
−
−
−
−
−
−
−
−
Table 4.3: C(2. M22 .2, 2F )
Class
2A
2DE
3A
4CD
4F
5A
6A
6BC
10A
11A
22A
4.4
∆1 (t)
1
5, 20
−
−
−
−
−
−
−
−
−
∆2 (t)
−
−
−
−
40, 40, 40, 40
160, 160
−
−
160, 160
−
−
∆3 (t)
−
−
40, 40
80
−
−
40, 40
80, 80
−
320, 320
320, 320
Leech Lattice Groups
The Leech Lattice is closely related to the Golay code, which is defined as follows. Let
Λ be a set of integral vectors (a1 , · · · , a24 ) such that:
• for i ∈ {1, 2, · · · , 24}, ai ≡ m mod 2;
4.4. Leech Lattice Groups
74
• The set {i : ai ≡ h mod 4} is in the Golay code, for each h, and
•
P24
i=1 ai
≡ 4m mod 8.
A group of 24-dimensional matrices will be the automorphism group of Λ, which has
central involution denoted by I24 . When we factor I24 out of the automorphism group
1 P24 2
a .
of Λ we get a simple group, Co1 . The Leech Lattice has a norm defined by
8 i=1 i
Therefore, the Conway groups Co2 and Co3 are defined by the Stabilizer of a vector
in Co1 with norm 4 and 6 respectively. To find the commuting involution graphs for
the double cover of the automorphism group of the Higman-Sims group, HS, and the
Suzuki group Suz we will use Algorithm 2. For the double cover of the Conway group
Co1 the situation is more complicated since the permutation rank in this case is 465,
and the permutation representation has 196560 points. This means, of course, the
inability to find the full CG (t)-orbits. For that reason we are going to build the next
algorithm which can be realised by appealing to Lemma 4.2.5 and Lemma 4.2.6 and
the fact that ∆i (t) is a union of specific CG (t)-orbits see (Lemma 4.2.1). Also parts 1
and 4 in Theorem 4.2.2 will be used in this algorithm to give the sizes of the suborbits
in ∆i (t).
Algorithm
3
.
Input: G is either 2.Co1 , t an involution in 2D;
i- for j ∈ {1, 2, · · · , ♯Classes(G)}; set ST Cj −→ 0; Repj −→ {},Xj −→ 0.
ii- there is k
∈
{1, 2, · · · , ♯Classes(G)} such that t is conjugate to
Classes(G)[k, 3].
iii- for i := 1 to ♯Classes(G) do
ST Cj −→ ST Cj + (CharacterT able(G)[i, j] ∗ CharacterT able(G)[i, k] ∗
Conjugate(CharacterT able(G)[i, k])) /(CharacterT able(G)[i, 1]).
iv- ST Cj 7−→ ST Cj ∗ Classes(G)[j, 2]/♯CG (t).
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
75
v- for x 7−→ Random(tG ) do
if t ∗ x
and
Classes(G)[j, 3]
conjugate in G, then
vi- Repj 7−→ Repj ∪ {x}; and Xj = Xj + (♯CG (t)/♯CCG (t) (x)), if ST Cj
Xj then Stop;
equal
else;
vii- pick y 7−→ Random(tG ) such that t ∗ y
and Classes(G)[j, 3] conjugate in
G, and no involutions in Repj are conjugate
with
y
in
CG (t) .
viii- repeat the steps[v,vi] until the step v holds.
ix- set N ewRep −→ {} , for x ∈ Repj and m ∈ divisor(order(t ∗ x)) do:
N ewRep 7−→ N ewRep ∪ {t ∗ (t ∗ x)m }
x- check the elements of N ewRep to see if they are in the same CG (t)-orbit, by
using IsConjgate Magma code.
xi- put REP to be the set of all CG (t)−orbit representatives, then Rep(∆1 (t)) 7−→
REP ∩ CG (t).
xii- for w ∈ REP Rep(∆1 (t)) check if CCG (t) (w) has an involution in tG then
w ∈ ∆2 (t); else;
xiii- search for w1 ∈ ∆1 (t), such that ∆1 (w1 ) ∩ ∆1 (w) 6= φ, therefore w ∈ ∆3 (t).
Output: the set of suborbit representative, ∆i (t) structure.
Now we utilise Algorithms 1 and 2 to obtain the disc structure of C(G, X), when G
is one of our aforementioned Leech Lattice Groups. In the following tables we give the
size of suborbits of CG (t)-orbits on X = tG inside the classes of G intersection ∆i (t):
• The first three tables deal with the groups 2. HS.2 and 2. Suz.2. In each case the
CG (t)-orbits are calculated and the disc structure is analyzed.
4.4. Leech Lattice Groups
76
Table 4.4: C(2. HS.2, 2B)
Class
2A
2CD
3A
4B
4D
6A
∆1 (t)
1
105
−
−
−
−
∆2 (t)
−
−
336
−
1260
336
∆3 (t)
−
−
−
56
−
−
Table 4.5: C(2. HS.2, 2E)
Class
2A
2CD
3A
4A
4B
4D
4F
5A
5B
5C
6A
6CD
7A
8D
10A
10B
10C
10DE
11A
12A
12B
14A
15A
20A
20B
20G
22A
30A
∆1 (t)
1
5, 60
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
∆2 (t)
−
−
160
−
−
60
240, 240
−
−
960, 960
160
480
−
−
−
−
960, 960
−
−
−
−
−
−
−
−
−
−
−
∆3 (t)
−
−
240
−
120
240, 240
−
−
320
−
240
480
1920
480, 480, 480, 480
−
320
−
960, 960
1920, 1920
960, 960
960, 960, 960, 960
1920
1920
960, 960, 960, 960
−
960, 960, 960, 960
1920, 1920
1920
∆4 (t)
−
−
−
240
−
240
−
192, 192
−
−
−
−
−
−
192, 192
−
−
−
−
−
−
−
−
−
1920
−
−
−
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
77
Table 4.6: C(2. Suz.2, 2D)
Class
2A
2BC
3A
3B
3C
4B
4CD
4E
4G
5A
5B
6A
6B
6CD
6E
6F G
6HI
7A
8KM
9A
10A
10B
10CD
11A
12BC
12D
12F
12H
12I
12K
14A
15B
18A
18B
18C
20EF G
22B
28A
30B
∆1 (t)
1
495
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
∆2 (t)
−
−
−
2640, 2640
2640
−
−
−
23760
−
47520
−
2640, 2640
−
2640
−
23760, 23760
−
−
−
−
47520
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
∆3 (t)
−
−
−
−
−
−
2970
23760
−
−
−
−
−
5940, 5940
−
15840, 15840, 15840, 15840
15840
47520
−
95040, 95040, 31680, 31680
−
−
47520, 47520
190080
23760, 23760
47520, 47520
47520, 47520
−
47520, 47520, 47520, 47520
47520, 47520, 47520, 47520
47520
95040, 95040
95040, 95040, 31680, 31680
95040, 95040, 95040, 95040
95040, 95040, 95040, 95040
95040, 95040
190080
190080
95040, 95040
∆4 (t)
−
−
132, 132
1760
−
1584
990
−
−
9504, 9504
−
132, 132
1760
3960, 3960
−
−
−
−
47520, 4750
−
9504, 9504
−
−
−
−
15840, 15840
15840, 15840
31680
−
47520, 47520, 47520, 47520
−
−
−
−
−
−
−
−
−
4.4. Leech Lattice Groups
78
• In the following table we compute the CG (t)-orbit sizes and their position in
∆i (t) for the commuting involution graph of 2. Co1 , with the involution class 2D.
Therefore, the structure of the graph is determined.
Table 4.7: C(2. Co1 , 2D)
Class
2A
2BC
2D
3A
3B
3C
3D
4A
4BC
4DE
∆1 (t)
1
7920, 495
95040, 1584, 25344
−
−
−
−
−
−
−
4F
−
4G
−
4H
−
5A
5B
5C
6A
6B
6C
6DE
−
−
−
−
−
−
−
∆2 (t)
−
−
−
−
135168, 135168, 112640
−
2027520
50688
15840, 126720
1520640,95040,95040,380160,
380160,760320
190080,190080,380160,3041280,
760320
190080,380160,380160,3041280,
253440,253440,253440,253440
1013760,1013760,6082560,
6082560
−
1622016, 1622016, 12165120
−
−
135168, 135168, 112640
−
−
∆3 (t)
−
−
−
2048
−
450560, 450560
−
−
−
−
−
−
−
1622016
−
9732096, 9732096
2048
−
450560, 450560
1013760
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
Class
6F
6GH
6IJ
∆1 (t)
−
−
−
6KL
−
6M N
6O
−
−
6P
7A
7B
8A
8CD
−
−
−
−
−
8E
8F
−
−
8GH
−
9A
9BC
−
−
10A
10B
10C
10DE
10F
−
−
−
−
−
10G
10H
−
−
10I
−
∆2 (t)
2027520
−
2027520,2027520,2027520,
2027520
2027520,2027520,2027520,
2027520,6082560,6082560,
1013760,4055040
−
6082560,6082560,6082560,
6082560,6082560,6082560,
6082560,6082560
24330240, 24330240
−
−
−
12165120,12165120,6082560,
6082560
12165120, 12165120
12165120,12165120,12165120,
12165120,12165120
6082560,6082560,6082560,
6082560,3041280,3041280,
3041280,3041280,12165120,
12165120,12165120,12165120
−
−
−
12165120,1622016, 1622016
−
−
24330240,24330240,24330240,
24330240
−
24330240,24330240,24330240,
24330240
−
∆3 (t)
−
4055040, 4055040
−
−
16220160, 16220160
−
−
16220160
32440320, 32440320
2027520, 2027520
−
−
12165120,12165120,
12165120,12165120
−
32440320, 32440320
32440320,32440320,32440320,
32440320
1622016
−
9732096, 9732096
24330240
−
48660480, 48660480
−
48660480, 48660480
79
4.4. Leech Lattice Groups
Class
10J
∆1 (t)
−
∆2 (t)
24330240,24330240,24330240,
24330240,24330240,24330240,
48660480
−
−
−
−
−
11A
12A
12BC
12D
12EF
−
−
−
−
−
12G
12HI
−
−
12J
12KL
−
−
12M
−
12N O
−
12P Q
−
12R
−
12S
13A
14A
14B
14CD
15A
15BC
15D
−
−
−
−
−
−
−
−
24330240,24330240,24330240,
24330240,12165120,12165120,
12165120,12165120
24330240,24330240,24330240,
24330240,24330240,24330240,
24330240,24330240,8110080,
8110080,8110080,8110080
97320960, 97320960
−
−
−
−
−
−
−
15E
−
−
16AB
−
−
18A
−
−
−
24330240,24330240,8110080,
8110080
48660480
−
24330240,24330240,24330240,
24330240,24330240,24330240,
24330240,24330240
−
80
∆3 (t)
−
194641920
3244032
1013760, 1013760, 6082560
24330240, 24330240
16220160,16220160,16220160,
16220160
48660480
8110080
−
48660480,48660480,24330240,
24330240
−
48660480,48660480,48660480,
48660480
−
−
−
97320960
16220160
32440320, 32440320
97320960, 97320960
19464192, 19464192
97320960
97320960, 97320960,
32440320, 32440320
97320960, 97320960,
97320960, 97320960
48660480, 48660480,
48660480, 48660480
32440320, 32440320
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
Class
18BC
18DE
18F GHI
20A
20B
20CD
20E
20F
∆1 (t)
−
−
−
−
−
−
−
−
20G
21A
21B
21C
22A
24CD
24E
24HIJ
26A
28A
28BC
30A
30BC
30DE
30F
30GHIJK
33A
35A
36AB
42A
42B
42C
52A
60AB
60CD
66A
70A
84A
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
∆2 (t)
−
−
−
48660480, 48660480
−
−
−
97320960,97320960,
97320960,97320960
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
81
∆3 (t)
32440320, 32440320, 32440320, 32440320
97320960, 97320960
97320960, 97320960, 97320960, 97320960
−
3244032, 3244032, 48660480, 48660480
48660480, 48660480
97320960, 97320960, 97320960, 97320960
−
97320960, 97320960, 97320960, 97320960
97320960
64880640, 64880640
194641920, 194641920
194641920
48660480, 48660480, 48660480, 48660480
32440320, 32440320, 32440320, 32440320
97320960, 97320960, 97320960, 97320960
97320960
194641920
97320960, 97320960, 97320960, 97320960
19464192, 19464192
97320960
97320960, 97320960
97320960, 97320960, 32440320, 32440320
97320960, 97320960, 97320960, 97320960
194641920, 194641920
194641920, 194641920
97320960, 97320960, 97320960, 97320960
97320960
64880640, 64880640
194641920, 194641920
194641920, 194641920
97320960, 97320960, 97320960, 97320960
194641920, 194641920
194641920, 194641920
194641920, 194641920
194641920, 194641920
4.5. Monster Sections
4.5
82
Monster Sections
In the 1970s Bernd Fischer during his research on a 3-transposition found what are
called now the Fischer groups F i22 , F i23 and F i′24 . They are defined by the finite
groups G satisfying the following :
• G is generated by a G-conjugacy class X with the property that if x1 , x2 ∈ X,
then order(x1 x2 ) 6 3;
• The derived subgroups G′ and G′′ are equal;
• If H 6 G such that H is a 2- or 3-subgroup, then H 6 Z(G).
Indeed, Fischer defined the commuting involution graph on X to prove that G′ is simple.
Now, According to Table 4.1 our aim is to construct the commuting involution C(G, X)
where
(G = 2.F i22 , X = 2F ), (G = 2.F i22 .2, X = 2B), (G = 2.F i22 .2, X = 2C)
and (G = 2.F i22 .2, X = 2F ). In fact, we are going to apply Algorithm 2 for the cases
(G = 2.F i22 .2, X = 2B), (G = 2.F i22 .2, X = 2C), while we will use Algorithm 3 to
calculate C(G, X) in the cases (G = 2.F i22 , X = 2F ) and (G = 2.F i22 .2, X = 2F ).
• The next table considers the structure of the commuting involution graph for
2.F i22 , with the class 2F . The calculations in this table involve CG (t)-orbit sizes.
Table 4.8: C(2. F i22 , 2F )
Class
2A
2BC
2DE
2F
3A
3B
3C
3D
4A
∆1 (t)
1
48
576, 216, 144, 9, 144
432,216,432,432,432,432,
216,6912,216,432,216
−
−
−
−
−
∆2 (t)
−
−
−
−
∆3 (t)
−
−
−
−
1536, 384, 384
8192
12288, 6144, 6144, 12288
49152, 49152
6912,576,6912,576,
9216,576,6912,6912,576,9216
−
−
−
−
−
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
Class
4BC
∆1 (t)
−
4D
4EF
−
−
4G
−
5A
6A
6B
6CD
6E
6F G
6H
6IJ
6KL
−
−
−
−
−
−
−
−
−
6M N
6OP
6Q
6R
−
−
−
−
6S
6T U
−
−
6V W
−
6XY
7A
−
−
8A
−
8B
−
∆2 (t)
6912,6912,6912,6912,
6912,6912
6912, 6912, 6912, 6912, 55296
55296,6912,6912,6912,55296,
6912,6912,6912
13824,13824,6912,6912,1152,
6912,13824,1152,13824,1152,
1152,6912,13824,6912,13824,
6912,6912,6912,13824,6912,
6912,13824,13824,6912,13824
,13824,6912,13824
−
384, 384, 1536
8192
4608, 4608, 4608, 4608
6144, 12288, 6144, 12288
24576
49152, 49152
73728
9216,9216,55296,6912,13824,
9216,9216,6912
36864, 36864, 36864, 36864
221184
110592, 110592
55296,55296,13824,55296,
13824,13824,13824,55296
110592, 110592
36864,55296,36864,36864,
55296,36864
110592,110592,110592,
110592
442368, 442368
442368,442368,147456,
147456
110592,55296,55296,55296,
110592,110592,55296,110592
18432,110592,18432,110592,
110592,110592,18432,18432
∆3 (t)
−
−
−
−
73728, 73728, 110592, 110592
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
83
4.5. Monster Sections
84
Class
8C
8DE
∆1 (t)
−
−
9A
9B
−
−
∆2 (t)
−
110592,110592,110592,
110592,110592,110592,
110592,110592,110592,
110592,110592,110592
−
−
9C
10A
−
−
−
−
10BCDE
−
−
12AB
12C
12D
−
−
−
12E
−
12F
−
12GHIJKLM N
12O
−
−
12P
12QR
13AB
14A
−
−
−
−
55296, 55296, 55296, 55296
73728, 73728, 73728, 73728
110592,55296,55296,55296,110592,
110592,110592,55296
55296,55296,55296,55296,
55296,55296,55296,55296
110592,110592,55296,55296,110592,
55296,110592,55296
221184, 221184, 221184, 221184
110592,110592,110592,
110592,110592,110592,
110592,110592,110592,
110592,110592,110592
221184,221184,221184, 221184
442368,442368,442368, 442368
−
−
14BC
−
−
15A
−
−
18A
18B
−
−
−
−
∆3 (t)
221184, 221184
−
147456, 147456
147456,147456,
442368, 442368
884736, 884736
73728,110592,
110592, 73728
221184,221184,
221184,221184
221184
−
−
−
−
−
−
−
−
884736, 884736
147456,442368,
442368,147456
442368,442368,
442368,442368
442368,442368,
442368,442368
147456, 147456
442368,442368,
147456,147456
Chapter 4.
Class
18GH
18IJ
18K
20AB
21A
24AB
26AB
30ABC
42A
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
∆1 (t)
−
−
−
−
−
−
−
−
−
∆2 (t)
−
−
−
−
−
−
−
−
−
85
∆3 (t)
442368, 442368, 442368, 442368
442368, 442368
884736, 884736
442368, 442368, 442368, 442368
884736, 884736
442368, 442368, 442368, 442368
884736, 884736
442368, 442368, 442368, 442368
884736, 884736
• In the final three tables we study the commuting involution graphs for the group
2.F i22 .2 , with the classes 2B, 2C and 2F . We end up with discs sizes.
Table 4.9: C(2. F i22 .2, 2B)
Class
2A
2DE
3A
6A
∆1 (t)
1
693
−
−
∆2 (t)
−
−
2816
2816
Table 4.10: C(2. F i22 .2, 2C)
Class
2A
2DE
3C
4D
6C
∆1 (t)
1
1575
−
−
−
∆2 (t)
−
−
22400
75600
22400
Table 4.11: C(2. F i22 .2, 2F )
Class
2A
2DE
2G
3A
3B
3C
∆1 (t)
1
27, 1080, 540
3240, 3240
−
−
−
∆2 (t)
−
−
−
2304
−
11520, 5760
∆3 (t)
−
−
−
−
5120
−
4.5. Monster Sections
86
Table 4.12: C(2. F i22 .2, 2F )
Class
4D
4F
4G
4I
4K
5A
6A
6B
6C
6JK
6LM
6T U
6V W
7A
8A
8E
8F
∆1 (t)
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
∆2 (t)
17280, 17280, 2160, 2160, 6480
103680, 51840
51840, 51840
51840, 51840
103680, 103680, 17280, 17280
414720, 27648
2304
−
5760, 11520
−
34560, 34560, 103680
103680
69120, 69120, 103680, 103680
−
207360, 207360
207360, 207360, 69120, 69120
−
9AB
10A
10CD
11A
12EF
12G
−
−
−
−
−
−
−
414720, 27648
414720, 414720, 414720, 414720
−
414720
−
12LO
12T
12U
14A
15A
16B
−
−
−
−
−
−
−
414720, 414720, 414720, 414720
−
−
−
−
18AB
18GH
22B
30A
−
−
−
−
−
−
−
−
∆3 (t)
−
−
−
−
−
−
−
5120
−
138240
−
−
138240
552960
−
−
414720, 414720,
829440
552960
−
−
3317760
−
138240, 138240,
138240, 138240
414720, 414720
−
829440, 829440
552960
1658880
1658880,1658880,
1658880,1658880
552960
1658880
3317760
1658880
Chapter 4.
Commuting Involution Graphs of C(2. G, X) and C(2. G.2, X)
87
The next theorem gives the structure of the commuting involution graph C(G, X),
where (G, X) is described in Table 4.1
Theorem 4.5.1. For G one of the groups of Table 4.1, we have the following results:
• DiamC(2.M12 .2, 2D) = 5
and
|∆1 | = 31, |∆2 | = 360, |∆3 | = 960, |∆4 | =
230, |∆5 | = 2.
• DiamC(2.M22 .2, 2F ) = 3 and |∆1 | = 51, |∆2 | = 800, |∆3 | = 1920.
• DiamC(2.Suz.2, 2D) = 4 and |∆1 | = 991, |∆2 | = 229680, |∆3 | = 3950100, |∆4 | =
536668.
• DiamC(2.HS.2, 2B) = 3 and |∆1 | = 211, |∆2 | = 1932, |∆3 | = 56.
• DiamC(2.HS.2, 2E) = 4
and
|∆1 | = 131, |∆2 | = 5660, |∆3 | = 37240, |∆4 | =
3168.
• DiamC(2.Co1 , 2D) = 3 and |∆1 | = 138799, |∆2 | = 2521524544, |∆3 | = 18839494656.
• DiamC(2.F i22 , 2F ) = 3 and |∆1 | = 12643, |∆2 | = 26511616, |∆3 | = 46448640.
• DiamC(2.F i22 .2, 2B) = 2 and |∆1 | = 1387, |∆2 | = 5632.
• DiamC(2.F i22 .2, 2C) = 2 and |∆1 | = 3151, |∆2 | = 120400.
• DiamC(2.F i22 .2, 2F ) = 3 and |∆1 | = 9775, |∆2 | = 9591984, |∆3 | = 29317120.
Proof. The proof of this theorem follows from Lemma 4.2.1 and the previous tables.
Chapter 5
Finite Groups Of Lie-type
In this chapter we give a summary about the finite groups of Lie-type, including the
construction and different widely known properties of such groups. This will put us in
good position for the next chapters, where we are going to investigate the commuting
involution graph of some of those groups. Finite groups of Lie-type such as the classical
matrix groups and their projective group may be defined in a variety of ways. Nevertheless, to completely realise the structure of these groups, it is beneficial to show them
in the scope of comprehensive theory. The method that we are going to consider is to
look at such groups algebraically.
5.1
Algebraic Group Theory
Now we will give a summary of some preliminary results related to algebraic groups.
As our purpose will be to quickly familiarize ourselves with those results we may utilize
in the coming chapters, we shall often full details of proofs and definitions may not be
provided. For the most accurate understanding of these topics we point to [43, 58].
Assume that k is an algebraically closed field. A subset V ⊆ k n ( the vector space on
n-tuples over k ) which is defined as the set of points which vanish on a finite set of
polynomial equations with coefficients in k is called an affine variety over k. The group
G is called an affine algebraic group if G is an affine variety and the maps
88
Chapter 5. Finite Groups Of Lie-type
89
G × G −→ G , (x, y) 7−→ xy; and
G −→ G , x 7−→ x−1
are morphisms of affine varieties. Furthermore, the Zariski topology on V is a topology space with closed sets defined to be the subvarieties of V . We should note that in
sections 5.1-5.4 we take k to be an algebraically closed field.
The typical example for G is the special linear group, SLn (k) which is defined as
2
SLn (k) = {(aij ) ∈ k n : det(aij ) − 1 = 0}.
Any closed subgroup of the general linear group GLn (k) is an affine algebraic group,
known as a linear algebraic group. As a matter of fact, it can be demonstrated that
every affine algebraic group is isomorphic to a linear algebraic group (and vice versa)
see [58], therefore we shall drop the term “algebraic group” to describe a linear or
affine algebraic group. Thus one can use matrices properties when looking at elements
of algebraic groups.
5.2
Subgroups of Algebraic Group
Let x ∈ End(V ) where V is a finite dimensional vector space over an algebraically
closed field k. Then x is semisimple if it is diagonalisable, and unipotent if the matrix
x has value 1 for all of its eigenvalues. The Jordan decomposition of these elements are
essential, and works as follows:
Theorem 5.2.1. [28] For x ∈ End(V ), x may be uniquely written as x = xs xu , such
that xs is semisimple, xu is unipotent and [xu , xs ] = Id(End(V )) .
Proof. See 1.4, [28]
Since an algebraic group G is a subgroup of GL(V ) for some V , so that it makes
sense to study the semisimple and it is unipotent elements of G. The unique maximal
5.3. Groups with a BN -pair
90
closed connected soluble normal subgroup of an algebraic group G is called the radical
of G and denoted by R(G). Furthermore, the unipotet radical of G, Ru (G) define to
be the unique maximal closed connected normal unipotent subgroup of an algebraic
group G. Moreover, G is reductive if Ru (G) = 1 and semisimple if R(G) = 1. Also
the algebraic group G is said to be simple if it has no proper closed connected normal
subgroups. Let k ∗ be the multiplicative group of k and let T be a subgroup of G such
that T ∼
= k ∗ × · · · × k ∗ . Then T is said to be a torus of G. The elements of T are
semisimple. In fact, for x a semisimple element of G there is a torus T of G containing
x. A maximal closed connected soluble subgroup of G is known as a Borel subgroup
of G. The next theorem establishes some principal results related to tori and Borel
subgroups of G:
Theorem 5.2.2. Let G be a linear algebraic group defined over an algebraically closed
field. The following hold:
1- All maximal tori in G are G-conjugate;
2- Any maximal torus lies in some Borel subgroup of G;
3- Any two Borel subgroups of G are G-conjugate;
4- If G is connected and B is a Borel subgroup of G then NG (B) = B.
Proof. See [43]
5.3
Groups with a BN -pair
Tits [66] presented the notion of a group with a BN -pair, and it has a vital role to play
in understanding the algebraic groups. Now we establish the main definition, followed
by the related results concerning algebraic groups.
Definition 5.3.1. A pair of subgroups B and N of a group G are called a BN -pair if
the following axioms hold:
1. B and N generate G;
Chapter 5. Finite Groups Of Lie-type
91
2. If K = B ∩ N , then K ⊳ N ;
3. The group W = N/K generate by a finite set of involutions S;
4. For ns ∈ N maps canonically to 1 6= s ∈ W , ns not normalise B;
5. For any s ∈ S and n ∈ N , we have ns Bn ⊆ Bns nB ∪ BnB.
The finite group W is called the Weyl group of G. Moreover, the pair (W, S) is
called a Coxeter system.
Now if G is a connected reductive linear algebra group. Associated with W is a
root system Φ = Φ+ ∪ Φ− with positive roots Φ+ and negative roots Φ− . Moreover,
corresponding to the root system Φ is a system of fundamental root {α1 , α2 , . . . αn }
which leads to the definition of the Dynkin diagram of G. For each root α ∈ Φ there
corresponds a subgroup (called a ”root subgroup” ) Xα of G. For further details in
that regard we refer the reader to ([28], 1.11).
On the other hand, a BN -pair of an algebraic group G is said to be a split BN -pair
if it satisfies the following conditions:
• B and N are closed subgroups of G;
• B = U (B ∩ N ) is a semidirect product of a closed normal unipotent group U and
a closed commutative subgroup B ∩ N , all of whose elements are semisimple;
• B∩N =
T
n∈N
nBn−1 .
One can prove that a connected reductive group G over algebraic closed field contains a split BN -pair (see [43], for instance). Now suppose that B is a Borel subgroup
of G and T is a maximal torus of G such that T ⊆ B. Thus if U = Ru (B), then B = U T
and B ∩ N = T . For clarity, suppose that G = SLn (k). Then if we let B to be the
upper triangular matrices, then B is Borel subgroup of G and the group of monomial
matrices is N . The subgroup of diagonal matrices is T and the upper uni-triangular
matrices form U .
5.4. Classification of Simple Algebraic Groups
5.4
92
Classification of Simple Algebraic Groups
The simple algebraic groups over an algebraically closed field k are classified by their
corresponding connected Dynkin diagram. There is an associated Dynkin diagram
to each connected reductive algebraic group G. If a Dynkin diagram of a connected
semisimple algebraic group is connected then G must be simple and of one of the types
presented in Figure 5.1. The group G has a unique Dynkin diagram but the converse
is not ture. For instance, the groups P GLn+1 (k) and SLn+1 (k) have the same Dynkin
diagram of type An . Full information regarding the Dynkin diagram may be found in
[28].
Al , l > 1
Bl , l > 2
Cl , l > 3
Dl , l > 4
E6
E7
E8
F4
G2
◦ − ◦ −··· − ◦ − ◦
α1
α2
αl−1 αl
◦ − ◦ −··· − ◦ ⇒ ◦
α1
α2
αl−1
αl
◦ − ◦ −··· − ◦ ⇐ ◦
α1
α2
αl−1
αl
◦α l
|
◦ − ◦ −··· − ◦ − ◦
α1
αl−2 αl−1
α2
◦α6
|
◦ − ◦ − ◦ − ◦ −◦
α1
α2
α3
α4 α5
◦α7
|
◦ − ◦ − ◦ − ◦ − ◦ −◦
α1
α2
α3
α4
α5 α6
◦α 8
|
◦ − ◦ − ◦ − ◦ − ◦ − ◦ −◦
α1
α2
α3
α4
α5
α6 α7
◦ −◦ ⇒ ◦ − ◦
α1 α2
α3
α4
◦⇛◦
α1
α2
Figure 5.1 Connected Dynkin diagrams
Chapter 5. Finite Groups Of Lie-type
5.5
93
Finite Groups of Lie-type
The previous sections were associated with algebraic groups over an algebraically closed
field. The finite group of Lie-type defined as a subgroups of connected reductive algebraic groups over an algebraic closed field of prime characteristic. This section will be
concerned with some basic concepts associated with finite groups of Lie-type. we start
with the abstract definition and then study some properties related to these groups.
Let k be the algebraic closure of the finite field Fp , where p > 0 is a prime. Let
G is a connected reductive group over k. Thus, G is isomorphic to closed connected
subgroup of GLn (k) for some n. Set q = pr for some r > 1 and define the map
Fq : GLn (k) −→ GLn (k) as follows
Fq : (aij ) 7−→ (aqij ).
Now, a homomorphism F : G −→ G, is called a standard Frobenius map if there
exist an injective homomorphism map i : G −→ GLn (k) for some n, satisfying
i(F (g)) = Fq (i(g)) for all g ∈ G and some q = pr .
Generally, A Frobenius map is a homomorphism F : G −→ G, with the condition that
F j , j > 1, is a standard Frobenius map. Moreover, if F is a standard Frobenius map
the fixed point group
GF = {g ∈ G : F (g) = g}.
is finite. If G is a connected reductive algebraic group and F is a Frobenius map the
finite group GF which arise in this way is called a finite group of Lie-type. A classification of finite, simple groups of Lie-type can be obtained from the classification of simple
algebraic groups. Nonetheless in some situations, the graph automorphism of Dynkin
diagrams together with an appropriate field automorphism yields extra simple groups,
when the field is finite. In the next chapter a complete list of finite simple groups of
Lie-type will be included.
5.5. Finite Groups of Lie-type
94
Now, suppose that G be a connected group over an algebraically closed field k of
characteristic p, and let F : G −→ G is an epimorphism such that GF is finite. Then
the map δ : G −→ G define as follows
δ(x) = x−1 F (x) for all x ∈ G.
is surjective. This result known as the Lang-Steinberg theorem has a vital role in
the theory of finite group of Lie-type see [28].
Let G be a connected reductive group with a Frobenius map F : G −→ G. Then
we aim to prove that G possess an F -stabilizer Borel subgroup. To prove that we take
any Borel subgroup B of G, as any two Borel subgroup are conjugate in G so that
B g = gBg −1 for some g ∈ G. Also, F (B g ) = B g if and only if F (B)g
−1 F (g)
= B. But
F (B) is also a Borel subgroup of G, thus there is h ∈ G such that F (B)h = B. Now by
applying Lang-Steinberg theorem we obtain h = g −1 F (g) for some g ∈ G. Therefore
B g is an F -stabilizer Borel subgroup of G. A Borel subgroup of GF is defined to be
B F , where B is an F -stabilizer Borel subgroup of G. One can prove that in a similar
way that any F -stabilizer Borel subgroup contains an F -stabilizer maximal torus. A
maximal tours of GF is defined to be a subgroup of the form T F , such that T is a
F -stabilizer maximal tours of G lies in an F -stabilizer Borel subgroup B F , for such
a torus T F , we get N = NG (T F ) is stable under F . Consequently, we obtain a split
BN -pair for GF from B F and N F .
Furthermore, it is not necessarily true that an F -stabilizer maximal torus of G lies
inside an F -stable Borel subgroup of G, so that not every maximal torus of GF lies
inside a Borel subgroup of GF . A maximally split is an F -stabilizer maximal torus of
G which lie in an F -stable Borel subgroup of G, whereas a torus of the form T F , such
that T is a maximal split torus of G is called maximal split torus of GF .
Chapter 5. Finite Groups Of Lie-type
95
Now we give an important results related to the Borel subgroup B F and maximal
torus of the group of Lie-type GF :
Theorem 5.5.1. For a finite group of Lie-type GF , the following properties hold:
1. Any two Borel subgroups of GF are GF -conjugate;
2. Any two maximal split torus subgroups of GF are GF -conjugate;
3. Any two maximal tori which lie in B F are B F -conjugate.
Proof. [43].
Chapter 6
Commuting Involution Graphs of
Exceptional Groups of Lie-type
6.1
Introduction
In the previous chapter, preliminary results relating to finite groups of lie-type were
displayed in abbreviated form. Our target in this chapter is to study the commuting
involution graphs, C(G, X), of a particular type of exceptional group of lie-type. One
can see from the classification of finite simple groups (see [31, 67] ) that exceptional
groups of Lie-type divide into two types:
• Untwisted groups: (An (q), n > 1), (Bn (q), n > 2), (Cn (q), n > 3), (Dn (q), n >
4), E6 (q), E7 (q), E8 (q), F4 (q), G2 (q).
• Twisted groups: (2 An (q), n > 1), (2 Dn (q), n > 1), 3 D4 (q), 2 E6 (q), (2 B2 (22n+1 )
= Sz(2n+1 ), n > 1), (2 G2 (32n+1 ) = Ree(32n+1 ), n > 1), (2 F4 (22n+1 ), n > 1), 2 F4 (2)′ .
Our investigation involves analyzing the disc structure and determining the diameters of the commuting involution graphs, C(G, X), when G is one of the following
groups:
• Untwisted groups : G2 (5), F4 (2), E6 (2).
• Twisted groups : R(27), 2 F4 (2)′ , 3 D4 (2), 3 D4 (3).
96
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
97
Let G is one of the groups above. Let t ∈ X where X is a conjugacy class of involution. To calculate the disc structure and the diameter for the commuting involution
graph C(G, X) we may use an approach similer to the one implemented in Chapter 3.
However, we may need in some cases to apply a new algorithm to determine C(G, X).
The permutation character of CG (t) in G has been found using the Magma code “PermutationCharacter” and hence the number of CG (t)-orbits (permutation rank on
X) under the action of X on CG (t) is calculated. Moreover, the complex character
tables of G are available in the GAP library. Therefore, we may use them to determine
the size of X and then use the Magma code “CentralizerOfInvolution” to compute
CG (t).
For our aforementioned groups the sizes of conjugacy classes of involutions and the
permutation ranks on each class of involution are listed in the next table.
Table 6.1: Disc sizes and Permutation Character
Group
G2 (5)
F4 (2)
F4 (2)
F4 (2)
F4 (2)
E6 (2)
E6 (2)
E6 (2)
R(27)
2 F (2)′
4
2 F (2)′
4
3 D (2)
4
3 D (2)
4
3 D (3)
4
Class
2A
2A
2B
2C
2D
2A
2B
2C
2A
2A
2B
2A
2B
2A
Size of Class
406875
69615
69615
4385745
350859600
5081895
8822169720
1587990549600
512487
1755
11700
819
68796
43584723
Permutation Rank
69
5
5
33
1002
5
62
719
54
5
30
4
27
209
6.2. Algorithms
98
One can note from Table 6.1 that many difficulties may arise during our study of
C(G, X) when G is F4 (2) and X = 2D, and when G is E6 (2) and X = 2C. The reason
for this is that these groups have large permutation rank on X, being respectively equal
to 1002 and 719 with big class size (see [67]). Consequently, we shall initialize a new
algorithm to determine the suborbits. However, Algorithm 2,3 which we established in
Chapter 4, will be employed to find the CG (t)-orbit on X for the remaining conjugacy
classes of involutions.
6.2
Algorithms
Let x ∈ X such that CCG (t) (x) ∩ X = ∅. By Theorem 4.2.2 part 4 we have that
d(x, t) > 3. In order to show that x ∈ ∆3 (t), it suffices to find an element y ∈ CG (x)
such that ty ∈ C, for some subset XC of ∆2 (t). Additionally, the Magma code “CentralizerOfInvolutuion” will be used to compute the centralizer of the involutions
in G and the code “RandomElementOfOrder” will be employed to find a random
element of given order. We apply the ideas above in the following algorithm to find
CG (t)-orbits in ∆3 (t):
Algorithm
4
.
Input: G is either E6 (2) or F4 (2), t an involution in 2C or 2D respectively;
i: set Rep(∆2 (t)) −→ the set of representative for C such that XC ⊆ ∆2 (t).
ii: for x 7−→ Random(X) do if CCG (t) (x) ∩ X = ∅ then
iii: repeat CInvolution −→ CentralizerOf Involutuion(G, x), in case G =
E6 (2) and CInvolution −→ Centralizer(G, x), for G = F4 (2).
iv: stop when the size of CInvolution equal to CG (t).
v: repeat y −→ RandomElementOf Order(CInvolution, 2)
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
vi: stop
when
#F ix(y)
equal
to
256
in
case
G
=
F4 (2)
99
or
Dimension(Eigenspace(y, 1)) equal 15 in case G = E6 (2).
vii: if order(t ∗ y) equal to order(z) for some z in Rep(∆2 (t)) then x ∈ ∆3 (t).
Output: the set of suborbit representatives in ∆3 (t) .
6.3
CG (t)-orbits
Untwisted Groups
Our aim is to calculate the commuting involution graphs for the untwisted groups
G2 (5), F4 (2) and E6 (2). The complex character tables for these groups are available,
therefore information about the number of conjugacy classes of involutions and the
size of XC can be acquired. For the groups G2 (5), F4 (2) we are going to use the
permutation representations with 3906 and 6988 points respectively from The Online
Atlas. Meanwhile, we consider E6 (2) as a matrix representation of dimension 27 over
GF (2). Moreover, from Table 6.1 we see that there is only one class of involutions
namely 2A in G2 (5). By calculation, for t ∈ 2A the size of the fixed-point set is 42. We
also see that F4 (2) contains four classes of involution, namely 2A, 2B, 2C and 2D with
size of fixed-point set equal to 5376 in 2A and 2B and 1280, 256 in 2C, 2D respectively.
Additionally, E6 (2) has three classes of involutions with dimension of fixed spaces being
respectively equal to 21, 17, 15 (information about fixed spaces will be available in the
next chapter as we use it to make a suborbit invariant). For a deeper understanding of
these groups we refer the reader to [27, 28, 71]. Now we provide tables giving the full
details about the CG (t)-orbits for t an involution in G2 (5), F4 (2) or E6 (2). Furthermore,
we include the sizes of the suborbits in each ∆i (t) belongs to.
6.3. CG (t)-orbits
100
Table 6.2: C(G2 (5), 2A)
Class
2A
3A
3B
4AB
5A
5B
5C
5D
5E
6AB
6C
7A
8AB
10AB
10CD
12AB
15A
15B
15E
20AB
21AB
24ABCD
25A
30AB
31ABCDE
∆1 (t)
200, 450
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
−
∆2 (t)
−
120
1200, 1800
1200, 1800
144
720
480, 1440
−
1440, 1440
1200, 1800
3600, 3600, 3600, 3600
−
7200, 7200
3600
7200, 7200
7200, 7200
1440, 1440
−
7200, 7200
7200, 7200
−
7200, 7200
−
7200, 7200
−
Table 6.3: C(F4 (2), 2A)
Class
2A
2C
3A
4C
∆1 (t)
270
2016
∆2 (t)
32768
34560
∆3 (t)
−
−
−
−
−
−
−
960
−
−
−
14400
−
−
−
−
−
14400
−
−
14400
−
14400
−
14400
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
Table 6.4: C(F4 (2), 2B)
Class
2A
2C
3A
4D
∆1 (t)
270
2016
∆2 (t)
32768
34560
Table 6.5: C(F4 (2), 2C)
Class
2AB
2C
2D
3AB
4AB
4CD
4F
4JK
4L
4M
5A
6GH
∆1 (t)
30
180,1920,1920,32,32
720,960,720,960,960,960,11520
∆2 (t)
32768
15360
11520
1024,1024
30720,30720
737280
184320,184320
1048576
983040
Table 6.6: C(F4 (2), 2D)
Class ∆1 (t)
2AB 192, 72, 3, 72, 12
2C
144,24,144,144,576,192,144,72,576,24,144,9,72
,12,144,576,144,72,192,72,576,12
2D
1152,144,576,576,144,144,576,144,144,1152,
1152,576,576,144,24,576,144,1152,24,576,144,
576,144,1152,1152,144,144,1152,144,144,1152,
576,144,576,144,576,576,576,144,144,576,576,
576,144,144,9216,144,576,1152,144,144,1152,
1152,24,144,24,1152,144,144,576,576,576,576,
1152,576,1152,576,144,1152,144,144,576,144,
1152
∆2 (t) ∆3 (t)
101
6.3. CG (t)-orbits
Class
3AB
3C
4AB
4CD
4EF
4GH
4I
4JK
4L
4M
∆1 (t)
102
∆2 (t)
2048,24576,6144
262144
576,12288,1152,576,9216,576,1152,
576,576,576,1152,1152,576,192,576
576,1152,1152,144,192,576,2304,576,
576,4608,576,144,4608,2304,576,288,4608,576,
576,2304,12288,576,288,192,288,576,576,288,
144,576,4608,192,2304,144,576
9216,9216,576,2304,1536,576,576,9216,1536,
9216,2304,576,9216,9216,1536,9216,1536,9216
,2304,2304
2304,4608,2304,2304,2304,4608,18432,4608,
9216,9216,4608,18432,18432,73728,4608,
18432,4608
18432,9216,18432,73728,9216,18432,9216,9216
,36864,18432,18432,9216,9216,18432,9216,
9216,73728,18432,9216,9216,9216,9216,36864
,9216,36864,36864,9216,18432
18432,9216,18432,4608,18432,4608,1152,
18432,18432,18432,4608,4608,18432,9216,
18432,18432,9216,18432,4608,4608,18432,
9216,9216,9216,2304,9216,18432,1536,
18432,18432,4608,18432,9216,1152,1152,
9216,4608,18432,1152,18432,18432,4608,
4608,4608,9216,9216,1536,4608,9216,4608,
1536,4608,9216,4608,9216,4608,2304,4608,
4608,18432,18432,1536,9216,18432,9216,
2304,18432,4608,2304,4608
36864,9216,9216,36864,36864,9216,36864,
9216,36864,9216,147456,9216,9216,9216,
36864,147456,36864,9216,36864
9216,18432,4608,9216,18432,18432,18432,2304
,4608,4608,18432,9216,36864,18432,2304,9216
,4608,9216,18432,9216,9216,9216,18432,18432
,9216,9216,9216,18432,18432,9216,18432,9216,
4608,9216,9216,4608,9216,18432,4608,18432,
9216,18432,9216,18432,18432,9216,18432,9216
∆3 (t)
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
Class
4M
4N
4O
∆1 (t) ∆2 (t)
18432,18432,4608,4608,18432,4608,18432,
18432,9216,18432,18432,4608,4608,9216,
18432,9216,9216,18432,18432,9216,18432,
9216,18432,18432,9216,36864,9216,18432,
9216,9216,18432,18432,18432,18432
147456, 147456, 147456, 147456
36864,36864,36864,36864,147456,36864,36864,
36864,147456,147456,36864,147456,36864,
36864,36864,36864
5A
6AB
6CD
6EF
6GH
6IJ
6K
7AB
8A
8B
8CDEF
8G
8HI
8J
103
∆3 (t)
196608, 196608,
589824
6144,24576,73728,73728,73728,6144,24576
36864,294912,73728,49152,36864,49152,73728 ,73728
786432
294912,73728,12288,49152,73728,73728,
147456,49152,147456,73728,36864,73728,
73728,73728,36864
73728,147456,73728,147456,73728,73728,
294912,73728,73728,73728,294912,
73728,294912,147456,147456,294912
2359296
1572864
294912,294912,294912,294912
294912,147456,147456,147456,147456,147456,
294912,294912,147456,147456,294912,147456
73728,73728,73728,73728,73728,294912,73728,
294912,147456,147456,147456,147456,73728,
73728,294912,73728,24576,73728,24576,294912
589824, 589824
294912,294912,294912,294912,294912,294912
294912,294912
589824,589824,589824,589824,589824,589824,
589824,589824,589824,589824,589824,589824,
589824,589824,589824,589824
6.3. CG (t)-orbits
Class
8K
104
∆1 (t) ∆2 (t)
589824,589824,589824,589824,
589824,589824
9AB
10AB
10C
12AB
12CD
12EF GH
12IJ
12KL
12M N
12O
13A
14AB
15AB
16AB
17AB
18AB
20AB
21AB
24ABCD
28AB
30AB
294912,294912,294912,294912
786432,786432
589824,294912,589824,589824,
98304,294912,294912,98304,
589824,294912
294912,294912,294912,294912,
294912,294912,294912,294912
294912,294912,294912,294912
∆3 (t)
4718592,1572864
589824,589824,1179648,1179648
589824,589824,1179648,1179648,
1179648,1179648
1179648
1179648
2359296,2359296
589824,589824,589824,589824,
589824,589824,589824,589824
589824,589824,589824,589824,
589824,589824
2359296,2359296,2359296,2359296
9437184
4718592
4718592,1572864
2359296,2359296,2359296,2359296
9437184
4718592,4718592
2359296,2359296,2359296,2359296
9437184
2359296,2359296,2359296,2359296
4718592,4718592
4718592,4718592
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
105
Table 6.7: C(E6 (2), 2A)
Class
2A
2B
3A
4B
∆1 (t)
2790
124992
−
−
∆2 (t)
−
−
2097152
2856960
Table 6.8: C(E6 (2), 2B)
Class
2A
2B
2C
3A
3B
4A
4B
4C
∆1 (t)
63,2160,2160
64512,120960,30720,4320,30240,30720,
30240,56
60480, 60480, 725760, 725760, 967680
−
−
−
−
−
4D
4E
4F
−
−
−
4J
4K
5A
6A
6D
6F
8C
12B
−
−
−
−
−
−
−
−
∆2 (t)
−
−
−
2359296
16777216
774144
967680,2211840,967680,2211840, 725760
3870720,3870720,3870720,1935360,
4423680,1935360,7741440,1935360,
7741440,1935360,4423680,4423680,
3870720,4423680
8847360,7864320,7864320
46448640,46448640
61931520,2064384,2064384,
61931520,61931520,61931520
123863040,123863040
743178240
939524096
70778880,70778880
990904320
1056964608
990904320,990904320
1132462080,1132462080
6.3. CG (t)-orbits
106
The final case is C(E6 (2), 2C). Here we shall make a different table for each ∆i (t),
because the size of the set XC is too large in most instances.
Table 6.9: ∆1 (t) Structure C(E6 (2), 2C)
Class
2A
2B
2C
∆1 (t)
3, 84, 1536, 2016
16128, 86016, 43008, 224, 5376, 168, 2016, 8064, 8064, 10752, 32256, 32256, 10752
258048,86016,64512,86016,16128,64512,64512,258048,36864,1032192,36864,36864,
32256,32256,64512,129024,16128,96,86016,129024,16128,32256,36864,32256,
258048,86016,129024,96,5376
Table 6.10: ∆2 (t) Structure C(E6 (2), 2C)
Class
3A
3B
3C
∆2 (t)
1572864, 917504
29360128
134217728
4A
4B
64512, 86016, 1032192, 64512, 32256, 786432, 21504, 36864, 36864, 64512, 36864, 1536
129024,1032192,1032192,64512,43008,258048,64512,129024,64512,32256,1032192,
32256,36864,786432,129024,86016,258048,32256,64512,1536,1536,16128,32256
516096,516096,1032192,516096,516096,2064384,64512,2064384,2064384,1032192,
516096,516096,4128768,129024,516096,516096,516096,2064384,258048,4128768,
129024,1032192,4128768,688128,64512,4128768,1032192,2064384,516096,129024,
1032192,64512,2064384,516096,129024,516096,129024,129024,258048,1032192,
2064384,688128,2064384,64512
4C
4D
2752512,1032192,2752512,11010048,1376256,16515072
4E
2064384,4128768,8257536,1032192,4128768,33030144,2064384,516096,
516096,516096,2064384,4128768,2064384,2064384,2064384,4128768,1032192,
4128768,4128768,258048,4128768,516096,2064384,4128768,4128768,4128768,
516096,4128768,2064384,258048,8257536,258048,2064384,4128768,4128768,
2064384,4128768,1032192,516096,1032192,2064384,1032192,1032192,4128768,
516096,1032192,1032192,4128768,516096,4128768,1032192,258048,4128768,
2064384,1032192,4128768,4128768,516096,4128768,4128768,516096
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
Class
4F
4G
4H
4I
4J
4K
6A
107
∆2 (t)
16515072,16515072,16515072,4128768,4128768,2752512,2752512,16515072,
5505024,33030144,33030144,33030144,16515072,5505024,2064384,1032192,
4128768,33030144,2064384,4128768,4128768,33030144,4128768,33030144,
16515072,16515072,4128768,16515072,1032192,4128768
8257536,66060288,16515072,16515072,8257536,33030144,16515072,33030144,
4128768,33030144,66060288,4128768,16515072
37748736,66060288,66060288,37748736
16515072,66060288,66060288,33030144,33030144,66060288,33030144,
33030144,16515072,33030144,66060288,264241152,66060288,66060288,
33030144,66060288,88080384,11010048,11010048
8257536,33030144,66060288,16515072,16515072,4128768,16515072,16515072,
16515072,8257536,8257536,8257536,66060288,66060288,33030144,66060288,
4128768,33030144,8257536,16515072,33030144,1376256,33030144,33030144,
33030144,8257536,4128768,33030144,33030144,66060288,66060288,16515072,
33030144,66060288,33030144,33030144,66060288,8257536,33030144,33030144,
33030144,33030144,33030144,33030144,8257536,2064384,33030144,33030144,
1376256,8257536,8257536,33030144,16515072,33030144,4128768,4128768,
66060288,8257536,66060288,2064384,16515072,4128768,8257536,8257536,
8257536,8257536,16515072,8257536
66060288,16515072,8257536,8257536,33030144,33030144,33030144,33030144,
66060288,66060288,264241152,8257536,16515072,33030144,16515072,16515072,
8257536,16515072,33030144,16515072,4128768,8257536,66060288,33030144,
66060288,33030144,16515072,33030144,66060288,33030144,66060288,33030144,
16515072,16515072,66060288,16515072,16515072,33030144,16515072,16515072,
8257536
44040192,44040192,33030144,33030144,2752512,33030144
6B
6C
6D
402653184
528482304,704643072
66060288,132120576,88080384,1835008,66060288,264241152,132120576,
88080384,132120576,528482304,66060288,66060288,176160768,132120576,
88080384
6E
132120576,88080384,37748736,132120576,88080384,528482304,
66060288,66060288,528482304,37748736
528482304,1056964608,88080384,352321536,704643072
2818572288
4227858432,1056964608,1056964608
8455716864
37748736,1572864,1572864,132120576,33030144,528482304,33030144,88080384,
37748736,132120576,66060288,66060288,88080384,528482304
6F
6G
6H
6I
8A
6.3. CG (t)-orbits
Class
8B
8C
8D
8E
8F
8G
8H
8I
8J
12A
12B
12C
12D
12E
12F
12G
12H
108
∆2 (t)
66060288,132120576,37748736,44040192,132120576,88080384,44040192,
66060288,528482304,88080384,37748736,528482304
132120576,264241152,33030144,264241152,264241152,66060288,66060288,
132120576,264241152,264241152,264241152,132120576,88080384,
88080384,132120576,66060288,66060288,264241152,132120576,
132120576,264241152,132120576,66060288,132120576,16515072,
66060288,66060288,264241152,132120576,66060288,264241152,
33030144,132120576,132120576,132120576,16515072
264241152,264241152,264241152,264241152,132120576,264241152,
264241152,264241152,264241152,264241152,264241152,132120576,
264241152,264241152,132120576,264241152,264241152,264241152,
264241152,264241152,264241152,264241152,1056964608,264241152,
132120576,1056964608
528482304,528482304,2113929216,176160768,2113929216,528482304,
528482304,176160768
2113929216,2113929216,2113929216,2113929216,2113929216
1056964608,528482304,1056964608,1056964608,528482304,528482304,
1056964608,264241152,1056964608,1056964608,1056964608,264241152,
264241152,1056964608,264241152,1056964608,528482304,1056964608,
1056964608,1056964608
2113929216,2113929216,2113929216
4227858432,4227858432,4227858432,4227858432,4227858432,4227858432
2113929216,2113929216,4227858432,2113929216,4227858432,4227858432,
1056964608,2113929216,1056964608,4227858432,2113929216,2113929216
402653184,402653184
132120576,528482304,132120576,528482304,528482304,528482304
1056964608,528482304,264241152,1056964608,528482304,264241152,
1056964608,264241152,264241152,1056964608,1056964608,1056964608,
1056964608,1056964608,264241152,528482304,1056964608,264241152,
1056964608,1056964608,1056964608,264241152,1056964608,1056964608,
1056964608,1056964608,264241152,528482304
1409286144,1409286144,2113929216,2113929216,4227858432,2113929216,
2113929216
4227858432,4227858432,4227858432
4227858432
5637144576
4227858432,2113929216,2113929216,2113929216,352321536,2113929216,
4227858432,4227858432,4227858432,352321536,2113929216,2113929216
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
Class
12I
12J
12K
12L
12M
12P
16A
16C
24A
∆2 (t)
8455716864, 8455716864
4227858432,2113929216,2113929216,1056964608,1056964608,
4227858432,4227858432,4227858432,4227858432,2113929216,
4227858432,2113929216,4227858432,2113929216,2113929216,
4227858432
4227858432,8455716864,8455716864,4227858432,1409286144,
1409286144,8455716864,4227858432,8455716864,4227858432
16911433728
16911433728,16911433728
16911433728,16911433728
8455716864,8455716864,8455716864,8455716864
16911433728,16911433728,16911433728,16911433728
8455716864,8455716864,8455716864,8455716864,8455716864,
8455716864,8455716864,8455716864
Table 6.11: ∆3 (t) Structure C(E6 (2), 2C)
Class
5A
7C
7D
9A
9B
10A
10B
12B
12E
12F
13A
14G
14H
15C
15D
17A
17B
18A
18B
20A
20B
∆3 (t)
1409286144,234881024
805306368
3221225472
22548578304
9663676416,3221225472
4227858432,2818572288, 2818572288
16911433728,8455716864, 2818572288,4227858432,4227858432
528482304,528482304
2818572288
8455716864
19327352832
16911433728
9663676416
22548578304
7516192768,22548578304
45097156608
45097156608
9663676416,9663676416
67645734912
16911433728,16911433728
33822867456,33822867456,33822867456,33822867456
109
6.3. CG (t)-orbits
110
Table 6.12: ∆3 (t) Structure C(E6 (2), 2C)
Class
21G
21H
24B
24C
24D
28K
28L
30E
30F
∆3 (t)
19327352832
45097156608
16911433728,16911433728,16911433728,16911433728
33822867456,33822867456
33822867456,33822867456
9663676416,9663676416
33822867456
22548578304, 22548578304
67645734912
Twisted Groups
The complex character tables for the groups R(27), 2 F4 (2)′ , 3 D4 (2) and 3 D4 (3) exist.
Therefore, the number of conjugacy classes of involutions and the size of XC are known,
both of which are very useful when we consider the commuting involution graph for
such groups. The permutation representation of R(27), 2 F 4(2)′ , 3 D4 (2) and 3 D4 (3)
come from (The Online Atlas) and the number of points is equal to 19684, 1600, 819
and 26572 respectively. However, from Table 6.1 we see that there is only one class
of involutions namely 2A in R(27). By calculation, for t ∈ 2A the size of fixed-point
set is 28. The Tits group 2 F4 (2)′ contains two classes of involution namely 2A and 2B
with size of fixed-point set is 0, 48 respectively. Additionally, 3 D4 (3) has one class of
involution with size of fixed space equal to 116. There are two classes of involution
in 3 D4 (2) with number of fixed spaces respectively equals 19, 27. A good resource
associated to these groups can be found in [27, 28, 71]. The next tables describe the
CG (t)-orbits for t an involution in R(27), 2 F4 (2)′ , 3 D4 (2) and 3 D4 (3). Furthermore, The
number of elements in each class is given.
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
Table 6.13: C(R(27), 2A)
Class
2A
3A
7A
9C
13ABCDEF
14ABC
19ABC
26ABCDEF
37ABCDEF
∆1 (t)
351, 351
∆2 (t)
∆3 (t)
728
4914,4914,4914,4914
19656
9828,9828
4914,4914,4914,4914
19656
9828,9828
19656
Table 6.14: C(2 F4 (4)′ , 2A)
Class
2A
2B
4C
5A
∆1 (t)
10
∆2 (t)
80
640
1024
Table 6.15: C(2 F4 (4)′ , 2B)
Class
2A
2B
3A
4A
4B
4C
5A
6A
8CD
12AB
13AB
∆1 (t)
12, 3
48,48,12,12,12
∆2 (t)
∆3 (t)
256,256
192,192
96,96
192,192,96
768
768,768
384,384
768,768
1536
111
6.3. CG (t)-orbits
112
Table 6.16: C(3 D4 (3), 2A)
Class
2A
3A
3B
3C
3D
4A
4B
6A
6B
6CD
7ABC
7D
8AB
9A
9B
12AB
13ABCDEF
13GHIJK
14ABC
21ABC
21DEF
26ABCDEF
28ABCDEF
28GHI
39ABCDEF
39GHIJKL
42ABC
52ABCDEF
56ABCDEF
73ABCDEF GHIJKLM N OP QR
78ABCDEF
84ABCDEF
104ABCDEF GHIJKL
∆1 (t)
2106,4536
∆2 (t)
∆3 (t)
224
8736
6048,8736
8736,11232
9072,8424
58968,117936
18144
235872
235872, 235872
19656
471744
235872, 235872
157248
471744
235872, 235872
19656
471744
58968,117936
78624, 78624
471744
58968,117936
58968,117936
117936, 117936, 117936, 117936
78624, 78624
471744
235872, 235872
235872, 235872
235872, 235872
471744
235872, 235872
235872, 235872
235872, 235872
Chapter 6. Commuting Involution Graphs of Exceptional Groups of Lie-type
Table 6.17: C(3 D4 (2), 2A)
Class
2A
3A
4A
∆1 (t)
18
∆2 (t)
∆3 (t)
512
288
Table 6.18: C(3 D4 (2), 2B)
Class
2A
2B
3A
3B
4A
4B
4C
6A
6B
7ABC
7D
8A
8B
9ABC
12A
13ABC
14ABC
18ABC
21ABC
28ABC
∆1 (t)
3,24
24,24,4,24,24,4,24,
4,24,4,24,24,24,4,24,
24, 4,24
∆2 (t)
∆3 (t)
384
512
24,24,24,24,192,24
24,24,24,24,192,24,24, 24,24,24,24
384,384,384,384,384,384
1536
384,384,384,384,384,384
512
3072
384,384,384,384,384,384
384,384,384,384,384, 384,384,384
512,1536
1536,1536
3072
1536
1536,1536
3072
1536,1536
113
6.4. Disc Structure
6.4
114
Disc Structure
The next theorem gives the structure of the commuting involution graph C(G, X),
where (G, X) is described in Table 6.1:
Theorem 6.4.1. For G one of the groups of Table 6.1, we have the following results:
• DiamC(G2 (5), 2A) = 3 and |∆1 | = 650, |∆2 | = 261264, |∆3 | = 144960.
• DiamC(F4 (2), 2A) = 2 and |∆1 | = 2286, |∆2 | = 67328.
• DiamC(F4 (2), 2B) = 2 and |∆1 | = 2286, |∆2 | = 67328.
• DiamC(F4 (2), 2C) = 2 and |∆1 | = 20944, |∆2 | = 4364800.
• DiamC(F4 (2), 2D) = 3 and |∆1 | = 50511, |∆2 | = 113896448, |∆3 | = 236912640.
• DiamC(E6 (2), 2A) = 2 and |∆1 | = 127782, |∆2 | = 4954112.
• DiamC(E6 (2), 2B) = 2 and |∆1 | = 2856311, |∆2 | = 8819313408.
• DiamC(E6 (2), 2C) = 3
and
|∆1 | = 3384671, |∆2 | = 609992912640, |∆3 | =
977994252288.
• DiamC(R(27), 2A) = 3 and |∆1 | = 702, |∆2 | = 314496, |∆3 | = 197288.
• DiamC(2 F4 (2)′ , 2A) = 2 and |∆1 | = 10, |∆2 | = 1744.
• DiamC(2 F4 (2)′ , 2B) = 3 and |∆1 | = 147, |∆2 | = 7712, |∆3 | = 3840.
• DiamC(3 D4 (2), 2A) = 3 and |∆1 | = 18, |∆2 | = 288, |∆3 | = 512.
• DiamC(3 D4 (2), 2B) = 3 and |∆1 | = 339, |∆2 | = 11112, |∆3 | = 57344.
• DiamC(3 D4 (3), 2A) = 3 and |∆1 | = 6642, |∆2 | = 27381536, |∆3 | = 16196544.
Proof. The proof of this theorem follows from Lemma 4.2.1 and the previous tables.
Chapter 7
Investigation On Commuting
Involution Graphs for the
Exceptional Groups of Lie-type
2E (2) and E (2)
6
7
7.1
Introduction
In this chapter we will study the commuting involution graphs for the exceptional
groups of Lie-type 2 E6(2) and E7(2) over GF (2). The factored order of 2 E6 (2) is
236 .39 .52 .72 .11.13.17.19 and the smallest matrix representation over GF (2) is of dimension 78. The smallest matrix representation over GF (2) for E7 (2) has dimension 56
and the factored order of E7(2) is 263 .311 .52 .73 .11.13.17.19.31.43.73.127. As we see
these groups have a large matrix representation dimension, besides large order. As a
result, the computational approach will resolve most of the calculations for such groups.
However, Magma commands which are used in the previous chapters to calculate the
size of suborbits or to distinguish between classes of involution will be unattainable.
In order to solve this problem we use the dimension of the fixed space and the size
115
7.2. Basic Definitions and Results
116
of the centralizer inside the maximal normal 2-subgroup of the centralizer in G of the
involutions to recognize different conjugacy classes of involutions. Moreover, many of
theoretical techniques may be applied to compute the sizes of the suborbits. Valuable
information about such groups can be found in [27, 49, 71].
7.2
Basic Definitions and Results
Let G be a finite group and t ∈ G be an involution. Set X = tG , a conjugacy class of
involutions. Our aim is to investigate the commuting involution graphs C(G, X) and
to analyze their disc structure for a finite simple groups.
Now we give the definition of the fix space, followed by a criteria to make a suborbit
invariant.
Definition 7.2.1. Let G be a matrix group and V be a corresponding G-module over a
finite field F of positive characteristic. Then the Fix Space of a random element g ∈ G
is defined as follows:
F ixg = {v ∈ V |vg = v}
It can be shown that F ixg is a subspace of V . It is also equal to the eigenspace of ρ(g)
associated to the eigenvalue 1 ∈ F, where ρ : G −→ GLn (F ) is a representation of G.
Lemma 7.2.2. Let G be a finite group and suppose that t ∈ X, where Xis a conjugacy
class of involutions in G. Then the following holds:
i- for x ∈ X we have F ixt ∼
= F ixx .
ii- for H ⊳ G and x ∈ X, we have StH (F ixt ) ∼
= StH (F ixx ).
iii- for x ∈ X and z ∈ xCG (t) , we have dimension(F ixt ∩ F ixx ) = dimension(F ixt ∩
F ixz ).
Proof. Since t is conjugate to x, there is h ∈ G such that th = hx. Now let
w ∈ F ixt , so that wh ∈ F ixx . Thus the map
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
117
Ψ : F ixt −→ F ixx ; w 7→ wh
is well defined. It is straightforward to check that Ψ is an isomorphism. Moreover, if
g ∈ StH (F ixt ) then h−1 gh ∈ StH (F ixx ).Therefore, the bijection
Υ : StH (F ixt ) −→ StH (F ixx ); g 7→ h−1 gh
is well defined and an isomorphism, so that [i] and [ii] are proved. Additionally, [iii]
follows from [i].
Let O2 be the maximal normal 2-subgroup of CG (t). The group O2 , has a substantial
role to play in examining whether the elements of X lie in same class under the action
of CG (t). The following lemma elucidates this:
Lemma 7.2.3. Let G be a finite group and let t ∈ X, where X is a conjugacy class
of involutions in G. Suppose that x, y ∈ X, such that xCG (t) = y CG (t) . For any normal
subgroup H of G the following holds:
i CH (x) ∼
= CH (y).
ii Let z ∈ H and Z = z CG (t) . We have |Z ∩ CH (x)| = |Z ∩ CH (y)|.
Proof. As xCG (t) = y CG (t) . We have x conjugate to y in CG (t), so there is h ∈ CG (t)
such that hxh−1 = y. Thus if w ∈ CH (x) then wh ∈ CH (y). For that reason the map
ς : CH (x) −→ CH (y), w 7→ wh
is isomorphism, and the prove of [i] is done. Part [ii] follows form [i].
7.3
The Commuting Involution Graphs of 2 E6 (2)
Let G be the 78-dimensional matrix representation over GF (2) of 2 E6 (2). GAP provides us with enough information about the number of conjugacy classes of involutions
and the size of XC as the character table of 2 E6 (2) is available. However, no details
about the permutation rank are available for this group. Moreover, the Magma code
7.3. The Commuting Involution Graphs of 2 E6 (2)
118
“IsConjugate” does not work. Therefore, alternative ways to deal with these difficulties will be established based on Lemma 7.2.2 and Lemma 7.2.3. The exceptional
group 2 E6 (2) has three classes of involution namely 2A, 2B and 2C with sizes respectively 3968055, 3142699560, and 1319933815200. The dimension of the fix space of these
classes is 56, 46 and 40 respectively. Here the study of the commuting involution graph
will be limited to the classes 2A and 2B. However, we shall show that C(2 E6 (2), 2C)
is connected. The full details about the C(2 E6 (2), 2A) and C(2 E6 (2), 2B) will be given
in the next tables, including CG(t)-orbits and discs structure.
Table 7.1: C(2 E6 (2), 2A)
Class
2A
2B
3A
∆1 (t)
1782
44352
−
∆2 (t)
4A
−
1824768
2097152
In contrast to C(2 E6 (2), 2A), the commuting involution graph C(2 E6 (2), 2B) is very
complicated. We apply Lemma 7.2.2 and Lemma 7.2.3 to distinguish between the
CG(t)-orbits. The method adopted for t ∈ 2B involves the following steps:
i- Figure out ∆1 (t) by working inside CG (t) and then using Theorem 4.2.2 part 2.
ii- Find the dimension of fix space of the CG (t)-orbits.
iii- Calculate O2 = pCore(CG (t), 2) and for z ∈ 2B find CO2 (z).
iv- Compute the CG (t)-classes in O2 ∩ ∆1 (t).
v- Check the position of the CG (t)-orbits by Algorithm 3 and Algorithm 6 (which
will be incorporated later in this chapter).
From step[iv] we see there are 6-classes of in O2 ∩ ∆1 (t), distributed as follows; one
class in X2A with size 63, in X2B there are 4-classes with size respectively 72, 4320,
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
119
4320 and 4320. Finally in X2C there is just one class with size 60480. Now for xi
C
a representative of this class and i = 1, 2, ..., 6 , let Ni = |xi G(t) ∩ CO2 (tCG (t) )|. The
following table gives the analyzes of C(2 E6 (2), 2B). For x ∈ X we give information
about the orbit xCG (t) including the size of the orbit and its position in the ∆i (t).
Details are given in the table below:
Table 7.2: C(2 E6 (2), 2B)
Class
2A
2B
2B
2B
2B
2B
2B
2C
2C
2C
3A
3B
4A
4B
4D
4E
4F
4H
4H
4L
4M
4M
5A
6C
6E
|F ix(t∗x) |
56
46
46
46
46
46
46
40
40
40
30
30
36
36
30
30
30
26
26
20
24
24
18
20
18
∆1 (t)
63
72
40320
4320
4320
4320
64512
60480
967680
60480
∆2 (t)
1835008
16777216
241920
1290240
8847360
8847360
8847360
2064384
2064384
247726080
123863040
123863040
1207959552
330301440
1056964608
|CO2 (x)|
16777216
16777216
262144
524288
524288
524288
16384
524288
16384
262144
1
1
262144
16384
256
256
256
16384
512
128
128
256
1
128
1
N1
63
63
15
63
63
63
31
63
15
15
0
0
20
12
0
0
0
16
0
0
0
0
0
0
0
N2
72
72
24
72
72
72
0
72
16
24
0
0
20
20
0
0
0
16
0
0
0
0
0
0
0
N3
4320
4320
672
2272
480
480
240
480
240
672
0
0
240
240
8
8
64
0
30
6
14
4
0
0
0
N4
4320
4320
672
480
2272
480
240
480
240
672
0
0
240
240
64
8
8
0
30
6
14
4
0
0
0
N5
4320
4320
672
480
480
2272
240
480
240
672
0
0
240
240
8
64
8
0
30
6
14
4
0
0
0
The diameter of the commuting involution graphs and the disc structure is shown
in the following next theorem. The above tables prove this theorem.
Theorem 7.3.1. Assume that G is the exceptional group of Lie-type 2 E6 (2) and X is
the conjugacy class of involution 2A or 2B in G. The discs structure of C(G, X) is as
N6
60480
60480
4224
6720
6720
6720
1440
8512
1440
4224
0
0
5520
1440
56
56
56
2160
180
60
36
124
0
78
0
7.3. The Commuting Involution Graphs of 2 E6 (2)
120
follows:
• if X = 2A then DiamC(G, X) = 2 and |∆1 | = 46134, |∆2 | = 3921920.
• if X = 2B then DiamC(G, X) = 2 and |∆1 | = 1206567, |∆2 | = 3141492992.
Finally, we will prove the connectivity of the C(2 E6 (2), 2C) as follows: The online
Atlas provides the 78-dimensional matrix representation of the group G = 2 E6 (2) such
that G = hx, yi where x has order 2 and y of order 3. Using a random search for an
involution t ∈ G whose fix space has dimension 40. Thus t ∈ 2C. Moreover, tx ∈ ∆1 (t)
and by apply Algorithm 6 (which will be incorporated later in this chapter) we get
ty ∈ ∆2 (t). Now for a ∈ CC there a path between t and a so that if we conjugate this
path by either x or y we get a new path between t and ax and the other between t
and ay . Consequently, x, y ∈ SG (CC ) and by Lemma 3.5.1 we obtain C(2 E6 (2), 2C) is
connected.
Chapter 7.
7.4
Commuting Involution Graph of 2 E6 (2) and E7 (2)
121
The Commuting Involution Graphs of E7 (2)
Assume that G is the 56-dimensional matrix representation over GF (2) of E7 (2). Let
t ∈ G be an involution and X = tG . As we have seen in Lemma 4.2.6 the character table
of G is very helpful during the investigation of CG (t)-orbits. However, the character
table of E7 (2) is unknown at the moment. In fact, since 1987 when Black and Fischer
calculated the complex character table of E6 (2), there has been no effort to determine
the full character table of the remaining exceptional groups. Consequently, we are
unable to compute the permutation rank of G on it is action on X or find the the size
of the set XC . Additionally, the GAP code“ClassMultiplicationCoefficient” cannot
be utilized. As the smallest matrix representation over GF (2) for G is of dimension
56, the attribute “IsConjugate” does not provide the the desired result. However,
the conjugacy classes of both involutions and semisimple elements of G are available.
In the next section we will present a statement relating to these classes. Moreover, the
dimension of the fix space of elements of even order will be applied to build a set of even
order elements with the same dimension of eigenspace. Utilizing the above information
we will be able to produce a subset of CG (t)-orbits and check its position in ∆i (t) of
C(G, X).
Involution of E7 (2)
Full details form this section can be found in [7], including a complete list of the
conjugacy classes of involutions of E7 (2). The involutions of E7 (2) are written in term
of the root system of type E7 . We supply the ordering of the root system of type E7
as follows:
7.4. The Commuting Involution Graphs of E7 (2)
122
Table 7.3: Ordering of the root system of type E7
α1
α2
α3
α4
α5
α6
α7
α8
α9
α10
α11
α12
α13
α14
α15
α16
α17
α18
α19
α20
α21
α22
α23
α24
α25
α26
α27
α28
α29
α30
α31
α32
α33
α34
(1
(0
(0
(0
(0
(0
(0
(1
(0
(0
(0
(0
(0
(1
(0
(0
(0
(0
(0
(1
(1
(0
(0
(0
(0
(1
(1
(0
(0
(0
(0
(1
(1
(1
0
1
0
0
0
0
0
0
1
0
0
0
0
0
1
1
0
0
0
1
0
1
1
0
0
1
0
1
1
1
0
1
1
0
0
0
1
0
0
0
0
1
0
1
0
0
0
1
1
0
1
0
0
1
1
1
0
1
0
1
1
1
1
0
1
1
1
1
0
0
0
1
0
0
0
0
1
1
1
0
0
1
1
1
1
1
0
1
1
1
1
1
1
1
1
2
1
1
1
2
1
1
0
0
0
0
1
0
0
0
0
0
1
1
0
0
0
1
1
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
1
0
0
0
0
0
1
1
0
0
0
0
1
1
0
0
0
1
1
1
0
1
0
1
1
1
0
1
1
0)
0)
0)
0)
0)
0)
1)
0)
0)
0)
0)
0)
1)
0)
0)
0)
0)
0)
1)
0)
0)
0)
0)
0)
1)
0)
0)
0)
0)
1)
1)
0)
0)
1)
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
α35
α36
α37
α38
α39
α40
α41
α42
α43
α44
α45
α46
α47
α48
α49
α50
α51
α52
α53
α54
α55
α56
α57
α58
α59
α60
α61
α62
α60 = α0
(0
(0
(1
(1
(1
(0
(0
(1
(1
(1
(0
(1
(1
(1
(0
(1
(1
(1
(1
(1
(1
(1
(1
(1
(1
(1
(1
(1
(2
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
1
1
2
1
2
1
2
2
2
2
1
1
2
1
1
1
1
2
1
1
1
2
2
1
1
2
2
1
2
2
2
2
2
2
2
2
2
3
3
2
1
2
2
1
2
2
2
2
2
2
2
2
2
2
3
2
2
3
3
2
3
3
3
3
3
4
4
4
1
1
1
1
1
2
1
1
2
1
2
2
1
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
1
1
0
1
1
1
1
1
1
1
1
1
1
1
2
1
1
2
1
1
2
1
2
2
2
2
2
2
2
123
0)
1)
0)
0)
1)
0)
1)
0)
0)
1)
1)
0)
1)
1)
1)
0)
1)
1)
0)
1)
1)
1)
1)
1)
1)
1)
1)
1)
1)
Now by identifying xαi with the root subgroup Uαi as defined in [7, 16.1]. Thus
we can obtain the structure of the conjugacy classes of involutions in E7 (2) as in the
following theorem:
Theorem 7.4.1. Let t ∈ E7 (2) be an involution , then t is of the form:
i- t1 = xα0 .
ii- t2 = xα59 xα58 .
iii- t3 = xα53 xα55 xα54 .
iv- t4 = xα48 xα47 xα49 .
7.4. The Commuting Involution Graphs of E7 (2)
124
v- t5 = xα53 xα49 xα47 xα48 .
Proof. See [7, 16.1] for full details of the proof.
The subgroup generated by removing the ith node from the Dynkin diagram is
a maximal parabolic subgroups of G denoted by Pi . In the next theorem we give a
relationship between the maximal parabolic subgroups of G and the centralizer of these
involutions.
Theorem 7.4.2. Let t1 , t2 , ...t5 the involutions shown in Theorem 7.3.1. There is a
maximal parabolic subgroup Pi of G containing the centralizer of these involutions as
described below:
i- CG (t1 ) 6 P1 ;
ii- CG (t2 ) 6 P6 ;
i- CG (t3 ) 6 P3 ;
i- CG (t4 ) 6 P7 ; and
i- CG (t5 ) 6 P2 , P7 .
Proof. See [7, 16.20] for full details of the proof .
Let Pi be a maximal parabolic subgroup of G, and denote the maximal normal
unipotent subgroup of Pi by Qi and its Levi complement by Li . The following theorem
sets up the structure of the centralizer of involution in G.
Theorem 7.4.3. The possible structure of CG (t), where t is an involution in G, is as
follows:
1. CG (t1 ) = Q1 L1 = P1 . Furthermore, Q1 is an extraspecial 2-group. Thus
Q1 /Z(Q1 ) has the structure of an orthogonal space, upon which L1 ∼
= Ω+
12 (2)
acts irreducibly;
2. CG (t2 ) = Q6 L, such that L ∼
= Sp8 (2) × Sym(3);
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
125
3. CG (t3 ) = CQ3 (t3 ) L, such that CQ3 (t3 ) 6 Q3 and L ∼
= Sym(3) × Sp6 (2) 6 L3 ;
4. CG (t4 ) = Q7 L, such that L ∼
= F4 (2) 6 L7 ;
5. CG (t5 ) = Q L, such that Q 6 Q2 Q7 and L ∼
= Sp6 (2) 6 L2 ∩ L7 ∼
= L6 (2).
Proof. see [7, 16.20]
In conformity with The Atlas giving priority to the larger centralizer, we may name
by 2A, 2B, 2C, 2D and 2E the conjugacy classes of the involutions t1 , t2 , t3 , t4 and t5
respectively. It can be shown in [13] that the dimension of the fix space of these
involutions is as follows:
Table 7.4: Dimension of Fix Space
Class
2A
2B
2C
2D
2E
dimF ix
44
36
32
28
28
The above result can be used as a tool to distinguish between different conjugacy
classes of involutions in G, with exception of the classes 2D and 2E, where they have
the same dimension on their fix space. However, we are going to apply the next lemma
to differentiate between the classes 2D and 2E.
Lemma 7.4.4. Let x and y be involutions of a finite group G . We have that x ∈ y G
if the order of x ∗ y is odd.
Proof. The basic proof can be seen from [10, 45.1,45.2].
Now let x and y be fixed involutions in the classes 2D and 2E respectively and let
z ∈ G be an involution with dimension of fix space equal to 28. In order to test whether
z ∈ 2D or z ∈ 2E we check the product x ∗ z g and y ∗ z g such that g runs as a random
element of G. Using lemma 7.4.4 we decide in which class z lies by taking the first odd
product.
7.4. The Commuting Involution Graphs of E7 (2)
7.4.1
126
Semisimple Classes of E7 (2)
Frank Luebeck in [49] showed that the group E7 (2) has 128 conjugacy classes of semisimple elements and his work included the order of the centralizer these elements. Moreover, Peter Rowley and his phd students (J.Ballantyne, C.Bates) in [13] provide the
structure of such centralizers. Full details about the semisimple elements of E7 (2) are
listed in the following table as shown in [13].
Table 7.5: Semisimple classes of E7 (2)
x
3A
3B
3C
3E
5A
7A
7B
7C
9A
9B
9C
9D
9E
9F
11A
13A
15A
15B
15C
15D
15E
15F
15G
17AB
19A
21A
21B
21CD
CG (x)
3.2 E6 (2).3
3 × Ω+
12 (2)
3 × U7 (2)
3. (U3 (2) × U6 (2)).3
5 × Ω−
8 (2) × Sym(3)
7 × L6 (2)
7 × 3 D4 (2)
7 × L3 (2) × L2 (8)
9 × 3 D4 (2)
9 × U5 (2)
9 × U3 (8)
9 × U4 (2) × Sym(3)
9 × L2 (8) × Sym(3)
[33 ].2.3.[22 ].[33 ].2
11 × 31+2 : 2.Atl(4)
13 × L2 (8)
15 × Ω−
8 (2)
15 × Alt(8) × Sym(3)
15 × Alt(8) × 3
15 × Alt(5) × (Sym(3))2
5 × 31+2
: 2.Alt(4) × Sym(3)
+
15 × Alt(5) × Sym(3) × 3
5 × 31+2
: 2.Alt(4) × 3
+
17 × Alt(5) × Sym(3)
19 × 9
21.L3 (4).3
21 × Alt(8)
21 × L2 (8)
Factored Order of CG (x)
236 .311 .52 .72 .11.13.17.19
230 .39 .52 .72 .11.17.31
221 .38 .52 .7.11.17
218 .310 .5.7.11
213 .35 .52 .7.17
215 .34 .5.73 .31
212 .34 .73 .13
26 .33 .73
212 .36 .72 .13
210 .37 .5.11
29 .36 .7.19
27 .37 .5
24 .35 .7
24 .37
23 .34 .11
23 .32 .7.13
212 .35 .52 .7.17
27 .34 .52 .7
26 .34 .52 .7
24 .34 .52
24 .35 .5
23 .34 .52
23 .35 .5
23 .32 .5.17
32 .19
26 .34 .5.72
26 .33 .5.72
23 .33 .72
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
x
21E
21F
21G
21H
31ABC
33AB
33CDEF GH
33I
35A
39A
43ABC
45AB
51AB
51CD
51EF
57AB
63ABC
63D
63EF G
63HIJ
73ABCD
85ABCD
91ABC
93ABC
99AB
105AB
105CDE
117ABC
127ABCDEF GHI
129ABCDEF
171ABCDEF
217ABCDEF
255ABCD
CG (x)
63 × L3 (2)
7 × 31+2
: 2.Alt(4)
+
21 × Alt(5) × Sym(3)
7×3×9
31 × L3 (2)
11 × 31+2 : 2.Alt(4)
11 × 31+2 : 2
11 × 31+2
35 × 3 × Sym(3)
117
43 × 3
5 × 9 × Sym(3)
51 × Alt(5)
51 × Sym(3)
51 × 3
19 × 9
63 × L3 (2)
63 × 7
63 × Sym(3)
63 × 3
73
85 × Sym(3)
91
93
99
35 × 3 × Sym(3)
35 × 3 × 3
117
127
129
171
217
255
Factored Order of CG (x)
23 .33 .72
23 .34 .7
23 .33 .5.7
33 .7
23 .3.7.31
23 .34 .11
2.33 .11
33 .11
23 .32 .5.7
32 .13
3.43
2.33 .5
22 .32 .5.17
2.32 .17
32 .17
32 .19
23 .33 .72
32 .72
2.33 .7
33 .7
73
2.3.5.17
91
93
32 .11
2.32 .5.7
32 .5.7
32 .13
127
3.43
32 .19
7.31
3.5.17
127
7.4. The Commuting Involution Graphs of E7 (2)
128
As we see, the exceptional group of Lie-type E7 (2) has 5 conjugacy classes of involution.
Our aim is to study the commuting involution graphs for the classes 2A, 2B and 2C by
finding a subset of XC with size equal or nearly equal to the size of XC . To deal with the
difficulties that arise during our search for CG (t)-orbits where t is an involution in one
of our aforementioned classes, we produce the algorithms given in the next subsection.
7.4.2
Algorithms
Let z ∈ G of order m and C = z G . The purpose of the first algorithm is to find a
subset of elements from the nonempty set XC . This algorithm is based on the results
of Lemma 7.2.2. We shall also note that the Magma code “UnipotentStabilizer”
and “Eigenspace” are involved. The algorithm is as follows:
Algorithm
5
.
Input: let t be an involution in E7 (2), and z ∈ G of order m and C = z G ;
i: d −→ Dimension(Eigenspace(z, 1)),O −→ M ultiples(m)∩ Divisors of |G|.
ii: repeat r −→ Random(tG ),o −→ order(t ∗ r).
iii: if o ∈ O, then s −→ IntegerRing()!(o/m) .
vii: stop when Dimension(Eigenspace((t ∗ r)s , 1)) equal d.
iv: ORB −→ {t ∗ (t ∗ r)s }
v: In order to get new orbits repeat the steps from ii to iv and let a2 be the
involution obtained in step iv and a1 ∈ ORB. They will be different if one of
the following holds
vi: #U nipotentStabilizer(O2 , Eigenspace(a1 , 1))) 6=
#U nipotentStabilizer(O2 , Eigenspace(a2 , 1)))
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
129
vii: Dimension(Eigenspace(t, 1) ∩ Eigenspace(a1 , 1) 6=
Dimension(Eigenspace(t, 1) ∩ Eigenspace(a2 , 1)
viii: ORB −→ ORB ∪ {a2 }
Output: the set of suborbit representatives with same order and same dimension of fix space.
As a matter of fact, for g ∈ G, the stabilizer of F ixg in G contains the centralizer in G of g. Therefore, we may apply this fact to check whether the elements
in CG(t) -orbits are in ∆2 (t). The Magma codes “Eigenspace”, “LMGSylow” and
“UnipotentStabilizer” are used in this work. The following algorithm serves this
purpose:
.Algorithm 6
Input: let t be an involution in E7 (2) and let b ∈ CG (t)-orbits, with order(t∗b) > 2;
i: E −→ (Eigenspace(b, 1)).
ii: Sy −→ LM GSylow(CG (t), 2),
U N −→ U nipotentStabilizer(Sy, E).
iii: if there is y ∈ U N such that (b, y) = (y, b) = Id(G) then b ∈ ∆2 (t).
iv: if not take Sy −→ Sy Random(CG (t)) and repeat the steps i and ii and check if
step iii holds.
Output: the set of suborbit representatives belonging to ∆2 (t).
The purpose of this algorithm is to locate a sufficient number of subgroups of the
centralizer in G of a semisimple element x ∈ G. The centralizer will then be the
subgroup generated by all such subgroups. The code “RandomElementOfOrder”,
7.4. The Commuting Involution Graphs of E7 (2)
130
“CompositionFactors” and “LMGcentralizer” contributed to the process of building this algorithm. We can also find the size of the CG (x) from Table 7.5. The following
algorithm is as follows:
.Algorithm 7
Input: let x be a semisimple element of G;
i: CG (x) = {Id(G)}, repeat z −→ RandomElementOf Order(G, 2).
ii: U −→ sub < GL(56, 2)|z, x >.
iii: M F −→ CopositionF actors(GM odule(U )).
iv: if size of M F greater than or equal to 20, then S −→ LM Gcentralizer(U, x).
iv: CG (x) −→ sub < GL(56, 2)|CG (x), S >.
vi: repeat the above steps until we have sufficient numbers of subgroups of
CG (x) to generate the centralizer of x in G.
Output: CG (x)
Definition 7.4.5. [32] Let G be a finite soluble group. A presentation for G of the
k
form < s1 , s2 , . . . , sn |sj j = hjj , 1 6 j 6 n, ssj i = hij , 1 6 i < j 6 n >, where
k
k
1- kj is the least prime such that sj j ∈< sj+1 , sj+2 , . . . , sn > for j < n, and sj j =
IdG if j = n, and
2- hij is a word in the generators for si+1 , si+2 , . . . , sn ,
will be called a power-conjugate presentation (pc-presentation) for G.
Now, a Sylow 2-subgroup of a finite group G contains representatives of all the
conjugacy classes of involutions of G. Thus we create the next algorithm to compute
a subset of CG (t)-orbits belonging to ∆1 (t). To do this we firstly need to calculate a
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
131
Sylow 2-subgroup of CG (t) by using the Magma code “LMGSylow(CG (t),2)”. We
then turn it into a pc-group using the code “LMGSolubleRadical”. This makes
it possible to obtain a set of representatives of a given class of involutions belonging
to the Sylow 2-subgroup of CG (t). By checking the action of CG (t) on this set by
conjugation we get the ∆1 (t). The Magma code “ConjugacyClasses”, “LMGIsIn”
and “LMGIsConjugate” are applied in this work. The algorithm is as follows:
Algorithm
8
.
Input: let t be an involution in E7 (2), C = CE7 (2) (t);
i: S −→ LM GSylow(C, 2).
ii: N1 , N2 , N3 −→ LM GSolubleRadical(S), N1 the soluble radical of S, N1 ∼
=
N2 ∼
= S, N2 is pc-group and N3 isomorphism from S to N2 .
iii: D −→ ConjugacyClasses(N 2).
iv: let K be a set of representative of conjugacy classes of involution in S. We
move them into representatives of conjugacy classes of involution in C.
v: a −→ Random(K).
vi: repeat c −→ Random(C).
vii: if LM GIsIn(S c , a) then T1 , T2 , T3 −→ LM GSolubleRadical(S c ).
viii: for x in K check if LM GIsConjugate(T2 , T3 (xc ), T3 (a)) then K −→ K \ {x}
Output: small set of representative of conjugacy classes of involution in
∆1 (t).
We are going to utilise all of the above algorithms, in addition to the ones in previous
chapters to investigate the commuting involution graphs.
7.4. The Commuting Involution Graphs of E7 (2)
7.4.3
132
Disks Structure and Orbit Size
For t be an involution in classes 2A, 2B or 2C in E7 (2). Let X = tE7 (2) , so that for
x ∈ X and m equal to order(t ∗ x), then by Theorem 4.2.2 part 2 if m = 2, then
x ∈ ∆1 (t). Moreover, if m is even, m > 4 and (t ∗ x)m/2 ∈ X, then x ∈ ∆2 (t). Otherwise we may employ Algorithm 6 to show that x ∈ ∆2 (t) or Algorithm 4 to prove
that x ∈ ∆3 (t). To calculate ∆1 (t) we first apply Algorithm 8 to get a small set of
representatives of involutions in CG (t), then we used Lemma 7.2.2 and Lemma 7.2.3 to
obtain a smaller set which allows us to break it into CG (t)-orbits by the Magma code
“Isconjugate”. Therefore we find the complete orbits of ∆1 (t).
On the other hand, to calculate the size of the orbit containing x it is enough to find
the CCG (t) (x). This is because the size of the CG (t)-orbit with representative x is equal
|CG (t)|
. However, to compute CCG (t) (x) we consider the following subcases :
to
|CCG (t) (x)|
1. If m = 2k and k is even, then we apply the following procedures to calculate
CCG (t) (x):
i- Compute the centralizer of the involution w1 = (t ∗ x)k inside CG (t) by using
“LMGCentralizer” Magma code. call this centralizer W1 .
ii- Calculate CW1 (w2 ),such the w2 = t ∗ x of order 2k. This can be done by
using the code “ApproximateStabilizer” to find the stabilizer in W1 for
the fix space of w2 namely StW1 (w2 ). We then apply the “OrbitBounded”
code to confirm the size of the stabilizer. Moreover, it is a straight forward
check to see that the generator set of StW1 (w2 ) commutes with w2 this leads
to CW1 (w2 ) = StW1 (w2 ). Set W3 = CW1 (w2 ).
iii- We find CW3 (x) by employing the codes “ApproximateStabilizer” and
“OrbitBounded” respectively. By Theorem 4.2.2 part 1 we have CCG (t) (x)
= CW3 (x).
2. If m = 2k and k is odd, then to find CCG (t) (x) we use the following steps:
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
133
i- Use Algorithm 7 to calculate the centralizer of the semisimple element w1 =
(t ∗ x)2 in G. Let W1 = CG (w1 ).
ii- Compute CW1 (w2 ) where w2 = t ∗ x using the “LMGRadicalQuotient”
code. We may let W3 = CW1 (w2 ).
iii- Again we employ the code “ApproximateStabilizer” to calculate CCG (t) (x)
= CW3 (x).
3. If m is odd, then in a similar way as case 2 we can calculate CCG (t) (x), but with
the advantage that we need only apply the parts ii and iii.
7.4.4
Connectivity
According to the online Atlas we may generate G = E7 (2) by two element x, y such
that x has order 5 and y has order 6. By Theorem 7.4.1 we know that G has five
G-conjugacy classes of involution. For t ∈ G an involution, to show the connectivity of
the graph we only need to prove that tx , ty ∈ CC . The reason for this is that for any
w ∈ CC there is path between t and w, so conjugating this path by x, y leads to two
new paths from tx to wx and from ty to wy . Here CC = StG (CC ) and the connectivity
of the graph follows by Lemma 3.5.1. The connectivity of C(E7 (2), 2A) follows by using
a random search to detect an involution t ∈ 2A such that tx , ty ∈ ∆1 (t). On the other
hand, one can find t ∈ 2B with properties tx ∈ ∆1 (t) and using Theorem 4.2.2 part 3
we can confirm that ty ∈ ∆2 (t). Hence C(E7 (2), 2B) is connected. Moreover, choosing
a random t ∈ 2C and applying Theorem 4.2.2 part 3 we get tx ∈ ∆2 (t) and Algorithm 4
shows that ty ∈ ∆3 (t). Therefore, the graph C(E7 (2), 2C) is connected. Finally, seeking
an involution t ∈ 2D or 2E and applying Algorithm 6 we get tx ∈ ∆2 (t). On the other
hand, there are x1 , x2 ∈ CG (ty ) conjugate to t satisfies (t, x1 ) = (x1 , x2 ) = (x2 , ty ) = 1.
Consequently, tx , ty ∈ CC and the connectivity of the graph is proved.
7.5. Commuting Involution Graph of 2A
7.5
134
Commuting Involution Graph of 2A
Let t be an involution in class 2A. We have the following:
1- t = xα0 .
2- The dimension of the fix space of t is 44.
3- The size of X = tG is equal to 18610317999.
4- CG (t) = P1 .
5- The 2-core of CG (t), namely O2 , has size 233 .
Our strategy is to compute the CG (t)-orbit by computing the set XC . This is because a non-empty set XC breaks down into CG (t)-orbits. Therefore, we are looking
for the CG (t)-orbit inside the non-empty set XC . Moreover, for x is a representative of
CG (t)-orbit, as O2 is a unipotent subgroup then the Magma code “UnipotentStabilizer”, is applicable to calculate the stabilizer of the fix space of x inside the 2-group
O2 , we name that by UO2 (x) and we used it as a CG (t)-orbit invariant (see Lemma 7.2.2
part ii). However, the code “ApproximateStabilizer” will be utilized to compute
the size of the orbits. We will now apply the above ideas as follows:
X2A By using Algorithm 5 there is an involution x ∈ X with t ∗ x ∈ 2A. We use the
code “ApproximateStabilizer” to find the Stabilizer in CG (t) for the fix space
of x, namely StCG (t) (x), and the “OrbitBounded” code to confirm the size of
the Stabilizer. By a straight forward check one can see that the generator set of
StCG (t) (x) commutes with x, and here CCG (t) (x) ∼
= StCG (t) (x). The size of this
orbit is equal to 151470 and |UO2 (x)| = 4294967296. Theorem 4.2.2 part 2 then
shows that x ∈ ∆1 (t).
X2B Algorithm 5 shows that there is x ∈ X such that t ∗ x ∈ 2B. In order to find
CCG (t) (x) we apply the “LMGRadicalQuotient” code on CG (t). This will give
an epimorphism from CG (t) to a permutation group isomorphic to the quotient
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
135
group CG (t)/O2 . Let W1 be the centralizer of the image of x and W2 to be the inverse image of W1 inside CG (t) so that CCG (t) (x) 6 W2 . The code “ApproximateStabilizer” provides us with StW2 (x) and by checking the generators of SW2 (x)
we obtain CCG (t) (x) ∼
= StW2 (x) and |StW2 (x)| = 4596373779694328217600. Additionally, using the above technique we get |UO2 (x)| = 33554432 and x ∈ ∆1 (t).
X3B By applying Algorithm 5 we obtain x ∈ X such that t ∗ x ∈ 3B. From table 7.5
and Algorithm 7 we calculate CG (t ∗ x) and by Theorem 4.2.2 part 1 we obtain
CCG (t) (x). We find that the size of this orbit is 8589934592 and |UO2 (x)| = 1.
Moreover, Algorithm 7 shows that x ∈ ∆2 (t).
X4 In this case Algorithm 5 gives an element x ∈ X with order(t ∗ x) = 4 and the
dimension of the fix space of t ∗ x is 32. To compute CCG (t) (x), let z = (t ∗ x)2 .
By checking the dimension of the fix space of z we obtain z ∈ 2A, and we
find CCG (t) (z) as above. Inside this group we apply the code “ApproximateStabilizer” to calculate CCG (t) (x). Therefore, we have the size of this orbit is
9926737920 and |UO2 (x)| = 65536, and x ∈ ∆2 (t) by Theorem 4.2.2 part 3.
From the above cases we see that a complete list of CG (t)-orbits are found, so that we
can establish the following theorem:
Theorem 7.5.1. Let G be the exceptional group E7 (2). Suppose that t ∈ 2A. We
have DiamC(E7 (2), 2A) = 2 with |∆1 | = 93645486 and |∆2 | = 18516672512.
Proof. From the above we can pin down the following table
Class Or Order
|F ix(t∗x) |
|UO2 (x)|
∆1 (t)
∆2 (t)
2A
44
4294967296
151470
−
2B
36
33554432
93494016
−
3B
32
1
−
8589934592
4
32
65536
−
9926737920
7.6. Commuting Involution Graph of 2B
7.6
136
Commuting Involution Graph of 2B
Let t be an involution in class 2B. We have the following:
1- t = xα59 xα58 .
2- The dimension of the fix space of t is 36.
3- The size of X = tG is equal to 6396887385160272.
4- CG (t) = Q6 L, where L = Sp8 (2) × Sym(3).
5- The 2-core of CG (t), namely O2 , has size 242 .
Computing the non-empty XC for the class 2B requiring a lot more time and effort
than we had in case the class 2A, besides we can not cover the complete list of CG (t)orbits. Thus we attempt as much as we can to collect the largest possible number of
CG (t)-orbits. Since no enough information about the conjugacy classes for the elements
of even order greater than 2, we let any two element of even order greater than 2 be
conjugate in G if they have the same order and the same dimension of fix space. For
x a random element in the set XC we calculate the size of the orbit represented by x
and determine which ∆i (t) it is located in.
C Is Involution Class
We will start first with the sets XC such that C is a conjugacy class of involution in G
where XC lies in ∆1 (t). For x an involution in such sets the code “LMGcentralizer”
will be employed to calculate its centralizer inside CG (t). This is possible because
x ∈ CG (t) and |CG (t)| is small enough to allow as to compute with it. Furthermore,
The method we employ to determine ∆1 (t) is established in section 7.4.3. This method
leads to the size of ∆1 (t), which is equal to 19113646671, and with 17 suborbits of
CG (t) . The next table explains the above statements:
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
137
Class
Dim(F ix(t∗x) )
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆1 (t)
2A
44
34
4398046511104
255
2A
44
30
137438953472
220320
2B
36
34
4398046511104
240
2B
36
29
137438953472
6609600
2B
36
26
4294967296
109670400
2B
36
27
8589934592
47001600
2B
36
26
67108864
53477376
2B
36
26
16777216
66846720
2B
36
26
3435973868
8812800
2C
32
26
4294967296
82252800
2C
32
26
3435973868
82252800
2C
32
24
16777216
4211343360
2C
32
24
536870912
5922201600
2D
28
26
67108864
50135040
2D
28
26
8589934592
50135040
2E
28
22
67108864
4211343360
2E
28
22
16777217
4211343360
7.6. Commuting Involution Graph of 2B
138
C Is Class of Order 3
Let x be an involution in 2B such that order(t ∗ x) = 3. There are just two such orbits
and they are in ∆2 (t) by Algorithm 6. The are as follows:
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆2 (t)
3B
24
131072
9126805504
3D
18
1
4398046511104
C Is Class of Order 4
We investigate the sets XC such that C is a class of elements of order 4. Now for x ∈ X
such that t ∗ x ∈ C Algorithm 6 shows that x ∈ ∆2 (t). By considering the dimension
of the fixed space for (t ∗ x), the CG (t)-orbits break up as follows:
(t ∗ x)2 − class
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2B
2B
2B
2B
2B
2C
2B
2B
Dim(F ix(t∗x) )
32
32
32
32
26
26
26
26
26
26
26
26
24
24
24
24
24
20
20
18
18
Dim(F ixt ∩ F ixx )
24
24
24
26
21
21
21
21
22
22
24
24
20
24
20
20
24
18
18
17
17
|UO2 (x)|
16777216
33554432
536870912
4294967296
16777216
2097152
8388608
1048576
8388608
67108864
536870912
33554432
1048576
131072
65536
1048576
16777216
65536
8192
65536
512
∆2 (t)
3743416320
14438891520
6768230400
658022400
14438891520
433166745600
433166745600
433166745600
126340300800
126340300800
13536460800
13536460800
6064334438400
8556380160
6930667929600
6930667929600
8556380160
43124156006400
172496624025600
6571299962880
6571299962880
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
139
C Is Class of Order 5
Suppose that x ∈ CG (t)-orbits such that order(t ∗ x) = 5. There is only one such orbits
and it is in ∆2 (t) by Algorithm 6, and is described as below:
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆2 (t)
5A
16
1
1055531162664960
C Is Class of Order 6
We now look at the sets XC such that C is a class of elements of order 6. Now for
x ∈ X such that t ∗ x ∈ C Algorithm 6 shows that x ∈ ∆2 (t). The set XC breaks up
into CG (t)-orbits according to the dimension of the fixed space as follows:
(t ∗ x)3 class
2A
2B
2B
2B
(t ∗ x)2 class
3B
3B
3B
3D
Dim
(F ix(t∗x) )
24
20
16
16
Dim(F ixt
∩F ixx )
20
18
16
16
|UO2 (x)|
∆2 (t)
65536
8192
512
1
7392712458240
229995498700800
7009386627072
1121501860331520
C Is Class of Order 8
Let x be in the set XC such that C is a class of elements of order 8 and dimension of
fix space equal 16. Then we have (t ∗ x)4 is conjugate to 2A , and Dim(F ix(t∗x)2 ) = 26.
In this case the CG (t)-orbits are described as follows:
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆2 (t)
16
256
1774250989977600
C Is Class of Order 12
In this case we are looking for the set XC such that C are a class of elements of order
12 and the dimension of the fix space is 16. For x ∈ X such that order(t ∗ x) = 12
7.7. Commuting Involution Graph of 2C
140
we have (t ∗ x)6 is conjugate to 2A , (t ∗ x)4 is conjugate to 3B, Dim(F ix(t∗x)3 ) =
32, Dim(F ix(t∗x)2 ) = 24 and x ∈ ∆2 (t) by Algorithm 6. To calculate CG (t) we use
Algorithm 7 to compute CCG (t) ((t ∗ x)6 ) and inside this group we find CG (t) by using
“ApproximateStabilizer” code. The CG (t)-orbits are as follows:
7.7
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆2 (t)
16
256
1892534389309440
Commuting Involution Graph of 2C
Let t be an involution in class 2C. We have the following:
1- t = xα53 xα55 xα54 .
2- The dimension of the fix space of t is 32.
3- The size of X = tG is equal to 26099300531453909760.
4- CG (t) = CQ3 (t3 ) L, such that CQ3 (t3 ) 6 Q3 and L = Sym(3) × Sp6 (2) 6 L3 .
5- The 2-core of CG (t), namely O2 , has size 245 .
Our target is to collect the largest subset of CG (t)-orbits for the C(G, 2C). Moreover,
the research involves determining the size of the CG (t)-orbits and the place of the orbits
in ∆i (t). Moreover, we note that Algorithms 3, 4, 5, 6, 7 are involved in this calculation
. We should note that the deal with sets XC such C is class of elements of even order
greater than 2 is the same as the case the class 2B. The set XC breaks up between the
CG (t)-orbits as described below:
C Is Involution Class
Let C be a conjugacy class of involution in G. For x in XC such set order(t ∗ x) = 2, we
have x ∈ ∆1 (t). Moreover, the procedure give in in Section 7.4.3 will cover all CG (t)orbits when C is an involution class. As a result of this we can figure out ∆1 (t). This
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
141
technique produces the size of ∆1 (t) which is equal to 49116616055 with 74 suborbits
of CG (t). Full information about these orbits is included in the following table:
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆1 (t)
2A
26
549755813888
120960
2A
26
2147483648
98304
2A
32
35184372088832
3
2A
28
35184372088832
1260
2B
26
35184372088832
15120
2B
26
35184372088832
16128
2B
26
549755813888
120960
2B
23
549755813888
8709120
2B
22
34359738368
30965760
2B
22
34359738368
23224320
2B
22
2147483648
33030144
2B
22
134217728
82575360
2B
22
8589934592
20643840
2B
22
2147483648
30965760
2B
23
68719476736
11612160
2B
24
68719476736
1032192
2B
24
549755813888
724760
7.7. Commuting Involution Graph of 2C
142
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆1 (t)
2C
20
17179869184
278691840
2C
20
2147483648
283115520
2C
20
2147483648
283115520
2C
20
2147483648
283115520
2C
20
2147483648
283115520
2C
20
2147483648
396361728
2C
20
2147483648
396361728
2C
20
2147483648
396361728
2C
20
2147483648
396361728
2C
20
67108864
2972712960
2C
20
67108864
2972712960
2C
20
67108864
990904320
2C
20
16777216
3963617280
2C
26
35184372088832
11520
2C
26
35184372088832
11520
2C
22
8589934592
92897280
2C
22
8589934592
247726080
2C
22
536870912
247726080
2C
20
536870912
990904320
2C
22
68719476736
69672960
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆1 (t)
2C
20
2147483648
1114767360
2C
22
134217728
247726080
2C
20
134217728
990904320
2C
20
2147483648
495452160
2C
20
17179869184
371589120
2C
22
549755813888
17418240
2C
22
68719476736
23224320
2C
22
549755813888
23224320
2C
20
4294967296
495452160
2C
20
4294967296
371589120
2C
20
17179869184
371589120
2C
20
17179869184
495452160
2C
24
549755813888
483840
2C
22
34359738368
23224320
2C
20
2147483648
371589120
2C
22
68719476736
30965760
2C
22
34359738368
30965760
2C
22
34359738368
30965760
2C
22
34359738368
30965760
2C
24
68719476736
1032192
2D
20
68719476736
15482880
2D
20
8589934592
15482880
2E
18
67108864
1981808640
143
7.7. Commuting Involution Graph of 2C
144
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆1 (t)
2E
18
536870912
1486356480
2E
18
16777216
11890851840
2E
18
268435456
1981808640
2E
18
268435456
1486356480
2E
20
4294967296
92897280
2E
20
68719476736
92897280
2E
20
67108864
2972712960
2E
20
67108864
99094320
2E
20
67108864
2972712960
2E
20
536870912
99094320
2E
20
536870912
743178240
2E
20
8589934592
743178240
2E
20
8589934592
557383680
2E
20
2147483648
557383680
C Is Class of Order 3
We consider the sets XC when C is a conjugacy class of elements of order 3. The
distribution of this set is as follows:
Class
Dim(F ixt ∩ F ixx )
|UO2 (x)|
∆2 (t)
3B
20
32768
6442450944
3B
20
16777216
2818572288
3E
14
1
35184372088832
3D
14
2048
5772436045824
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
145
C Is Class of Order 4
We now look at the sets XC where C is a conjugacy class of elements of order 4 which
have the same dimension of fix space. Full information about the size of these orbits
and their position in the ∆i (t) is drawn up in the following table:
(t ∗ x)2 − class
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2B
2B
2B
2A
2B
2B
2A
2A
2A
2A
Dim(F ix(t∗x) )
32
32
32
32
32
26
26
26
26
26
26
26
26
26
26
26
26
26
24
24
24
24
24
24
24
24
24
22
Dim(F ixt ∩ F ixx )
20
20
20
20
22
17
17
17
17
17
17
18
18
18
18
19
19
20
16
16
16
16
16
16
16
17
18
16
|UO2 (x)|
16777216
536870912
33554432
32768
549755813888
8388608
4194304
33554432
67108864
16777216
134217728
33554432
536870912
2147483648
134217728
2147483648
134217728
4294967296
524288
524288
1948576
32768
8388608
2097152
4194304
16777216
134217728
4194304
∆2 (t)
3963617280
3963617280
3963617280
3221225472
3221225472
285380444160
285380444160
142690222080
142690222080
142690222080
23781703680
23781703680
17836277760
17836277760
23781703680
4459069440
11890851840
743178240
2029372047360
2029372047360
2029372047360
2029372047360
2283043553280
2283043553280
1141521776640
285380444160
23781703680
1141521776640
7.7. Commuting Involution Graph of 2C
(t ∗ x)2 − class
2B
2B
2B
2B
2B
2C
2C
2B
2B
2C
2B
2B
2B
2C
2B
2B
2B
2B
2B
2C
2C
2C
2C
2C
2C
2C
Dim(F ix(t∗x) )
20
20
20
20
20
20
20
20
20
20
20
20
20
20
18
18
18
18
18
16
16
16
16
16
16
16
146
Dim(F ixt ∩ F ixx )
14
14
14
14
14
14
14
14
14
15
16
16
18
18
13
13
13
13
14
12
12
12
12
12
13
14
|UO2 (x)|
16384
32768
32768
524288
1048576
512
65536
2097152
131072
524288
8388608
131072
134217728
67108864
32768
65536
2097152
262144
4194304
16384
8192
4096
4096
512
65536
524288
∆2 (t)
24352464568320
24352464568320
16234976378880
16234976378880
6088116142080
64939905515520
32469952757760
1522029035520
18264348426240
18264348426240
1141521776640
1014686023680
31708938240
42278584320
48704929136640
48704929136640
2029372047360
2029372047360
1522029035520
97409858273280
97409858273280
292229574819840
292229574819840
389639433093120
73057393704960
24352464568320
C Is Class of Order 5
We investigate the set XC such that C is a conjugacy class of elements of order 5. We
only found two such orbits and details about them are listed in the following table:
Class
5A
5A
Dim(F ixt ∩ F ixx )
12
12
|UO2 (x)|
64
2048
∆2 (t)
1108307720798208
277076930199552
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
147
C Is Class of Order 6
We examine the sets XC with C a conjugacy class of elements of order 6, with the
property that they have the same dimension of the fix space. For an orbit representative
x in such sets we will determine the size of the orbit of x and it is place within the
∆i (t). Details are given in the table below:
(t ∗ x)3 class
2C
2C
2C
2C
2C
2A
2A
2B
2B
2C
2C
2A
2B
2B
2B
2B
2B
2B
2A
2A
2A
2A
(t ∗ x)2 class
3E
3D
3D
3D
3D
3D
3D
3D
3D
3B
3B
3E
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
Dim
(F ix(t∗x) )
12
12
12
12
12
16
16
16
16
16
16
20
20
20
20
20
20
20
24
24
24
24
Dim (F ixt ∩
F ixx )
10
10
10
10
12
12
12
12
14
12
12
14
14
14
14
14
14
15
16
16
16
18
|UO2 (x)|
∆2 (t)
1
64
128
512
1024
1024
64
1204
2048
512
16384
1
512
32768
65536
16384
524288
2097152
2097152
65536
32768
16777216
33249231623946240
16624615811973120
12468461858979840
12468461858979840
1039038488248320
519519244124160
554153860399104
1039038488248320
17317308137472
389639433093120
129879811031040
105553116266496
129879811031040
97409858273280
43293270343680
43293270343680
24352464568320
18264348426240
3044058071040
2164663517184
2029372047360
42278584320
C Is Class of Order 7
We analyze the set XC such that C is a conjugacy class of elements of order 7. For x
in such a set the information on the orbits is included in the following table:
Class
7A
7C
Dim(F ixt ∩ F ixx )
14
8
|UO2 (x)|
1
1
∆3 (t)
211106232532992
10133991615836160
7.7. Commuting Involution Graph of 2C
148
C Is Class of Order 8
Here we check the sets XC with C a conjugacy class of elements of order 8 having the
same dimension of fix space. For an element x in the set, we will determine the size of
the orbit of x and the position of x in the ∆i (t), as shown in the following table:
(t ∗ x)4 class
2C
2C
2C
2C
2B
2B
2B
2B
2B
2B
2B
2C
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
Dim
Dim
Dim
(F ix(t∗x) ) (F ix(t∗x)2 ) (F ixt ∩
F ixx )
8
16
8
8
16
8
10
20
9
10
20
9
12
20
10
12
20
10
12
20
10
12
20
10
12
20
10
12
20
10
12
20
12
12
20
10
14
26
11
14
26
11
14
26
11
14
24
11
16
26
12
16
26
12
16
26
12
16
26
12
16
26
12
16
26
12
20
32
14
20
32
14
20
32
14
20
32
14
|UO2 (x)| ∆2 (t)
16
32
256
32
32
512
64
1024
128
256
8192
2048
512
2048
8192
4096
32768
16384
4096
512
8192
131072
1024
512
8192
65536
14962542307758080
4987384735919360
4987384735919360
4987384735919360
12468461858979840
12468461858979840
12468461858979840
12468461858979840
9351346394234880
9351346394234880
779278866186240
18702692788469760
4675673197117440
2337836598558720
2337836598558720
779278866186240
58445914963960
58445914963960
58445914963960
259759622062080
389639433093120
146114787409920
129879811031040
129879811031040
129879811031040
32469952757760
Chapter 7.
Commuting Involution Graph of 2 E6 (2) and E7 (2)
149
C Is Class of Order 9
In this case we study the set XC such that C is a conjugacy class of elements of order
9. For x a representative on such an orbit important details are listed in the following
table:
(t ∗ x)-class
9D
9E
(t ∗ x)3 −class
3E
3E
Dim(F ixt ∩ F ixx )
10
8
|UO2 (x)|
1
1
∆3 (t)
709316941310853312
303992974847508480
C Is Class of Order 10
We regard the sets XC with C a conjugacy class of elements of order 10 with equal
dimension of fix space. For an orbit representative x in these sets we calculate the size
of the orbit of x and the location of x in the ∆i (t), as shown in the table below:
(t ∗ x)5 -class
2C
2C
2A
2B
2B
2B
Dim(F ix(t∗x)
8
8
12
12
12
12
Dim(F ixt ∩ F ixx )
8
8
10
10
10
10
|UO2 (x)|
16
32
512
32
64
128
∆2 (t)
199495389743677440
66498463247892480
8312307905986560
66498463247892480
13299692649578496
16624615811973120
C Is Class of Order 12
We discuss the sets XC with C a conjugacy class of elements of order 12 with equal
dimension of fix space. For an element x in such sets we calculate the size of the orbit
of x and the location of x in the ∆i (t). This is summarized in table below:
7.7. Commuting Involution Graph of 2C
(t ∗ x)6 class
(t ∗ x)4 class
2B
2B
2C
2B
2B
2A
2A
2B
2B
2B
2B
2B
2B
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
2A
3E
3D
3D
3B
3B
3D
3E
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3B
3E
150
Dim
Dim
Dim
Dim
(F ix(t∗x) ) (F ix(t∗x)2 ) (F ix(t∗x)3 ) (F ixt ∩
F ixx )
8
12
20
8
8
16
20
8
8
12
20
8
10
20
18
9
10
20
18
9
12
16
32
10
12
20
24
10
12
20
20
10
12
20
20
10
12
20
20
10
12
20
24
10
12
20
20
10
12
20
24
10
14
24
26
11
14
24
26
11
14
24
26
11
14
24
26
11
14
24
26
11
14
24
26
12
14
24
26
12
14
24
26
12
16
24
32
12
16
24
32
12
16
24
32
12
16
24
32
12
16
24
32
12
20
20
32
14
|UO2 (x)| ∆2 (t)
1
32
16
32
256
64
1
512
32
256
256
128
1024
512
2048
4096
8192
1024
16384
131072
32768
32768
4096
131072
16384
512
1
398990779487354880
66498463247892480
398990779487354880
4987384735919360
16624615811973120
22166154415964160
33249231623946240
12468461858979840
24936923717959680
37405385576939520
18702692788469760
18702692788469760
12468461858979840
4675673197117440
4675673197117440
4675673197117440
4675673197117440
4675673197117440
779278866186240
58445914963960
58445914963960
779278866186240
779278866186240
2922295748119840
86586540687360
259759622062080
105553116266496
C Is Class of Order 13
We contemplate the set XC such that C is a conjugacy class of elements of order 13.
Information about the CG (t)-orbit of x can be seen in the table below:
Class
13A
Dim(F ixt ∩ F ixx )
8
|UO2 (x)|
1
∆3 (t)
607985949695016960
Commuting Involution Graph of 2 E6 (2) and E7 (2)
Chapter 7.
151
C Is Class of Order 14
We consider the sets XC with C a conjugacy class of elements of order 14 and having
the same dimension of fix space. For an orbit representative x in such sets we compute
the size of the orbit of x and in which ∆i (t) the element x sitting. The following table
describes our findings:
(t ∗ x)7 class
(t ∗ x)2 class
Dim
(F ix(t∗x) )
2A
2B
7C
7A
8
12
Dim
(F ixt ∩
F ixx )
8
10
|UO2 (x)| ∆3 (t)
1
1
303992974847508480
66498463247892480
C Is Class of Order 15
We search for the sets XC such that C is a conjugacy class of elements of order 15. For
x in such sets, we provide essential details on the orbits in the following table:
(t ∗ x)-class
15A
15B
(t ∗ x)5 -class
3B
3D
Dim(F ixt ∩ F ixx )
8
10
|UO2 (x)|
1
1
∆3 (t)
425590164786511872
70931694131085312
C Is Class of Order 16
We looking for the sets XC with C a conjugacy class of elements of order 16 with equal
dimension of fix space. For an element x in such sets we determine the size of the orbit
of x and its position in the ∆i (t), as described in the following table :
(t ∗ x)8 class
2B
2B
2A
Dim
Dim
Dim
Dim
(F ix(t∗x) ) (F ix(t∗x)2 ) (F ix(t∗x)4 F ixt ∩
F ixx )
8
12
20
8
8
12
20
8
8
14
20
8
|UO2 (x)|
|∆3 (t)|
4
8
16
398990779487354880
398990779487354880
299243084615516160
7.7. Commuting Involution Graph of 2C
152
C Is Class of Order 17
We investigate the sets XC such that C is a conjugacy class of elements of order 17.
For x in such sets, details about the orbits are listed in the following table:
(t ∗ x)-class
17A
Dim(F ixt ∩ F ixx )
8
|UO2 (x)|
1
∆3 (t)
851180329573023744
C Is Class of Order 18
We explore the sets XC with C a conjugacy class of elements of order 18 having the
same dimension of fix space. For an orbit representative x in such sets we calculate the
size of the orbit of x and determine in which ∆i (t) it is sitting, as shown in the table
below:
(t ∗ x)9 class
(t ∗ x)6 class
(t ∗ x)2 class
Dim
F ix(t∗x)3
2A
3E
9E
20
Dim
F ixt ∩
F ixx )
8
|UO2 (x)|
∆3 (t)
1
303992974847508480
C Is Class of Order 20
We now look at the sets XC with C a conjugacy class of elements of order 20 with
equal dimension of fix space. For an element x in such sets we compute the size of the
orbit of x and identify the ∆i (t) containing this orbit. Our findings are summarized in
the table below:
(t∗x)10 class
Dim
F ix(t∗x)
Dim
F ix(t∗x)2
Dim
F ix(t∗x)5
2A
2B
2B
8
8
8
12
12
12
32
24
24
Dim
F ixt ∩
F ixx )|
8
8
8
|UO2 (x)|
∆2 (t)
16
8
4
132996926495784960
53198770583139840
53198770583139840
Commuting Involution Graph of 2 E6 (2) and E7 (2)
Chapter 7.
153
C Is Class of Order 21
In this case we investigate the set XC such that C is a conjugacy class of elements of
order 21. For x in such orbits, noteworthy details are recorded in the following table:
(t ∗ x) -class
(t ∗ x)3 class
(t ∗ x)7 class
21G
21D
7A
7C
3D
3B
Dim
(F ixt ∩
F ixx )
8
8
|UO2 (x)|
∆3 (t)
1
1
851180329573023744
607985949695016960
C Is Class of Order 24
we examine the sets XC with C a conjugacy class of elements of order 24 with dimension
of fix space equal 8. We notice that for an orbit representative x in such sets the
dimension(F ixt ∩ F ixx ) = 8. The following table describes information about this type
of set:
(t ∗ x)8
-class
3B
3D
3B
(t∗x)12
-class
2A
2A
2B
Dim
F ix(t∗x)6
26
32
20
Dim
F ix(t∗x)4
24
16
20
Dim
F ix(t∗x)3
16
20
12
Dim
F ix(t∗x)2
14
12
12
|UO2 (x)| ∆i
16
18
4
Orbit Size
∆2 (t) 299243084615516160
∆2 (t) 53198770583139840
∆3 (t) 398990779487354880
C Is Class of Order 28
We look at the sets XC with C a conjugacy class of elements of order 28 with dimension
of fix space equal 8. For an element x in such sets the dimension(F ixt ∩ F ixx ) = 8. In
the following table we include details on such orbits :
(t∗x)14 class
2A
2B
(t ∗ x)4 class
7C
7A
Dim(F ix(t∗x)7 )
Dim(F ix(t∗x)2 )
|UO2 (x)| ∆3 (t)
32
20
8
12
1
1
303992974847508480
797981558974709760
7.7. Commuting Involution Graph of 2C
154
C Is Class of Order 30
We consider the sets XC with C a conjugacy class of elements of order 30 with dimension of fix space equal 8. One can check that for an orbit representative x in
such sets the dimension (F ixt ∩ F ixx ) = 8, dimension(F ix(t∗x)3 ∩ F ixx ) = 12 and
dimension(F ix(t∗x)6 ∩ F ixx ) = 16. Information on these orbits can be found in the
following table:
(t∗x)1 5 class
2A
2B
(t∗x)10 class
3B
3D
Dim
F ix(t∗x)5
24
16
Dim
F ix(t∗x)2
8
12
|UO2 (x)| ∆i
Orbit Size
1
1
425590164786511872
1063975411966279680
∆2 (t)
∆3 (t)
Chapter 8
Conclusions and Future Work
We end this thesis by providing a brief summary of our work and presenting a view as
to potential future related research for the interested reader.
8.1
Conclusion
In conclusion, we mention that this thesis is centred on analyzing the commuting involution graph of finite group G, such that G belongs to one of the following groups:
i- The double cover of the symmetric group Sn ;
ii- The double cover and the automorphism group of the double cover of some sporadic groups;
iii- Some exceptional groups of Lie-type.
Many results have been achieved throughout this work. For instance, the connectivity and disc structure of such graphs were determined. However, establishing the
diameters of these graphs is the most prominent of what has been accomplished. A variety of methods have been employed for this purpose along the lines of the traditional
theoretical style as is the case in the Chapter 3. On the other hand, in the remaining
chapters computational approaches were most applicable in studying the commuting
involution graphs.
155
8.2. Potential for Future Work
8.2
156
Potential for Future Work
The results that have been obtained in this thesis open the door to examine several
interesting trends in future work:
1. The completion of the study of the commuting involution graphs for the groups
E7 (2) and 2 E6 (2). In addition, investigating the commuting involution graphs for
more complicated groups like E8 (2).
2. Calculating the collapsed adjacency matrix for C(G, X), which is defined as follows: Let t ∈ X, set Or1 , Or2 ...Orn be a complete list of CG (t)-orbits of X and
for i = 1, 2, .., n pick a representative ti ∈ Ori . Then the n × n matrix with
(i, j)th entry the value |Orj ∪ ∆1 (xi )| is called the collapsed adjacency matrix for
C(G, X).
3. The π-product graph Pπ (G, X) is the graph with vertex set X a non-empty
subset of a finite group G with two distinct vertices x, y ∈ X connected by an
edges if the product xy has order in π (a subset of the natural numbers). Valuable
information about these graphs can be found in [70]. Let X be a conjugacy class
of involution. If π = {2}, the P2 (G, X) is a commuting involution graph and
if π is the set of odd natural numbers then Pπ (G, X) is the local fusion graph,
F(G, X). This graph was studied deeply in ([14, 15]). In the future we offer the
reader to study the π-product graph Pπ (G, X) in the case that X is a conjugacy
class of involution, π is the set of all even natural numbers greater than 2 and G
is one of the groups mentioned in this thesis.
Appendix A
Magma Implementations
The following Magma implementations were employed to study the commuting involution graphs. However, various functions have been created for this purpose. The
first one is called Orbits and aims to calculate the discs structure of the graph and the
second one computes the structure constants and is denoted by StructureConstants.
The last one called XCE which aims to find a random element of the set XC .
r,t:=RandomElementOfOrder(G,2);
C:=CentralizerOfInvolution(G,t);
B:=Conjugates(G,t);
repeat;
z:=Random(B);
until z ne t;
Reps:={z};
D:=B diff {z,t};
Conj:=Conjugates(C,z);
time
for x in D do if x notin Conj then
Conj:=Conj join Conjugates(C,x);
Reps:= Reps join {x};
157
158
end if;end for;
function Orbits(S1,S2);
S:={};
for x in S1 do for y in S2 do
if Order(x*y) eq 2 then
Include(~S,x);
end if;end for;end for;
return S;
end function;
RDisc1:=Orbits(Reps,{t});
Disc1:={};for x in RDisc1 do
Disc1:=Disc1 join Conjugates(C,x);end for;
R1:=Reps diff RDisc1;
if #R1 ne 0 then
RDisc2:=Orbits(R1,Disc1);end if;
Disc2:={};for x in RDisc2 do
Disc2:=Disc2 join Conjugates(C,x);end for;
R2:=R1 diff RDisc2;
if #R2 ne 0 then
RDisc3:=Orbits(R2,Disc2);end if;
Disc3:={};for x in RDisc3 do
Disc3:=Disc3 join Conjugates(C,x);end for;
CL:=Classes(G);
CT:=CharacterTable(G);
function StructureConstants(x,y)
D:=0;
for i:=1
to
#CT do
D:=D+(CT[i,x]*CT[i,y]*Conjugate(CT[i,y]))/(CT[i,1]);
end for;
Appendix A.
Magma Implementations
return (D*CL[x,2])/ #Centralizer(G,CL[y,3]);
end function;
function XCE(t,m);
O:={};for x in Divisors(#G) do if not IsDivisibleBy(x, m) then
Include(~O,x);end if;end for;
s:=Id(G);
repeat;
x:=t^Random(G);
o:=Order(t*x);
if o notin O then
s:= IntegerRing()!(o/m);
end if;
until Order((t*x)^s) eq m;
return (t*(t*x)^s);
end function;
159
Appendix B
Magma Codes
Here we provide the most used Magma codes in our work, giving their functions as
described in the online handbook [32]
Magma Code
Random
ApproximateStabilizer
Description
A randomly chosen element for the group G
A is image of representation of G and A acts on U , a subspace or vector. Approximate the Stabilizer of U under A.
CentralizerOfInvolution
Given an involution g in G, this function returns the Centralizer C of g in G using an algorithm of John Bray.
CharacterTable
Construct the table of irreducible characters for the group
G.
Eigenspace
The eigenspace of the matrix a, corresponding to the eigenvalue e.
IsConjugate
Given a group G and elements g and h belonging to G,
return the value true if g and h are conjugate in G.
Given a matrix group G defined over a finite field, the intrinsic returns the Centralizer in the matrix group G of g ∈ G.
Given a matrix group G defined over a finite field, the intrinsic returns the conjugacy classes of G.
Given a matrix group G defined over a finite field, this intrinsic returns the factored order of G.
LMGCentralizer
LMGConjugacyClasses
LMGFactoredOrder
160
Appendix B. Magma Codes
161
Magma Code
LMGIsIn
Description
Given a matrix group G defined over a finite field Fq , and
an element x of the generic over group GL(n, q) of G; if
x ∈ G then the intrinsic returns true and the corresponding
element of WordGroup(G); and false otherwise.
LMGRadicalQuotient
Given a matrix group G defined over a finite field, the intrinsic returns a permutation group P isomorphic to G/L,
where L is the soluble radical of G. An epimorphism G to
P and its kernel L are also returned.
LMGSolubleRadical
Given a matrix group G defined over a finite field, the intrinsic returns the soluble radical S of G. A group P of type
GrpPC and an isomorphism S −→ P are also returned.
Given a matrix group G with natural module M and an
object y which is either a vector of M , a submodule of M , or
a tuple whose components are either vectors or submodules,
return true if the orbit of y under G has length less than or
equal to b. Otherwise the function returns false.
Given a group G and some subgroup H of G, construct
the ordinary character of G afforded by the permutation
representation of G given by the action of G on the coset
space of the subgroup H in G.
A randomly chosen element for the group G
Given a finite matrix group G, this intrinsic attempts to
locate an element x of order n in G by random search
Given a group G and a prime p, construct a Sylow psubgroup of G.
OrbitBounded
PermutationCharacter
Random
RandomElementOfOrder
Sylow
UnipotentStabilizer
Given a unipotent subgroup G of GL(d, F ), for F a finite
field, U a subspace of the natural vector space, determine
the Stabilizer in G of U .
Bibliography
[1] A. Abdollahi, S. Akbari, H. Dorbidi and H. Shahverdi. Commutativity pattern of
non-abelian p-groups determine their orders. Comm. Algebra, 41:451-461, (2013).
[2] A. Abdollahi, S. Akbari, and H. R. Maimani. Non-commuting graph of a group. J.
Algebra, 298(2):468-492, (2006).
[3] A. Abdollahi. Commuting graphs of full matrix rings over finite fields. Linear Algebra Appl., 428(11-12):2947-2954, (2008).
[4] S. Akbari, A. Mohammadian, H. Radjavi, and P. Raja. On the diameters of commuting graphs. Linear Algebra Appl., 418(1):161-176, (2006).
[5] S. Akbari and P. Raja. Commuting graphs of some subsets in simple rings. Linear
Algebra Appl., 416(2-3):1038-1047, (2006).
[6] M.Aschbacher and Y.Segev. The uniqueness of groups of Lyons type. J. Amer. Math.
Soc., 5(1):75-98, (1992).
[7] M. Aschbacher and G. M. Seitz. Involutions in Chevalley groups over fields of even
order. Nagoya Math. J. 63:191, (1976).
[8] M. Aschbacher. A condition for the existence of a strongly embedded subgroup. Proc.
Amer. Math. Soc., 38:509-511, (1973).
[9] M. Aschbacher. 3-transposition groups. Volume 124 of Cambridge Tracts in Mathematics. Cambridge University Press, Cambridge, (1997).
162
Bibliography
163
[10] M. Aschbacher. Finite group theory. Volume 10 of Cambridge Studies in Advanced
Mathematics. Cambridge University Press, Cambridge, second edition, (2000).
[11] J. Araújo, M. Kinyon, and J. Konieczny. Minimal Paths in the Commuting Graphs
of Semigroups. European J. Combin., 32(2):178-197, (2011).
[12] L. Babai and E. Szemeredi. On the complexity of matrix group problems I. Foundations of Computer Science, Annual IEEE Symposium on, 0:229-240, (1984).
[13] J. Ballantyne, C. Bates and P. Rowley. The Maximal Subgroups of E7 (2). LMS J.
Comput. Math. 18(1):323-371, (2015).
[14] J. Ballantyne, N. M. Greer and P. J. Rowley. On coprimality graphs for symmetric
groups. Graphs and Combinatorics., 29(6):1595-1622, November (2013).
[15] J. Ballantyne On local fusion graphs of finite coxeter groups. J. Group Theory,
16:595-617, (2013).
[16] C. Bates, D. Bundy, S. Hart and P. Rowley. Commuting involution graphs for
sporadic simple groups. J. Algebra, 316(2):849-868, ( 2007).
[17] C. Bates, D. Bundy, S. Hart, and P. Rowley. A note on commuting graphs for
symmetric groups. Electron. J. Combin., 16(1): Research Paper 6, 13, (2009).
[18] C. Bates, D. Bundy, S. Perkins and P. Rowley. Commuting involution graphs for
finite Coxeter groups. J. Group Theory, 6(4):461-476, (2003).
[19] C. Bates, D. Bundy, S. Perkins and P. Rowley. Commuting involution graphs for
symmetric groups. J. Algebra, 266(1):133-153, (2003).
[20] C.Bates, D. Bundy, S. Perkins and P. Rowley. Commuting involution graphs in
special linear groups. Comm. Algebra, 32(11):4179-4196, (2004).
[21] R. Brauer, and K. A. Fowler. On Groups of Even Order. Ann. of Math., Vol. 62,
no. 3:565-583, (1955).
Bibliography
164
[22] J. N. Bray. An improved method for generating the centralizer of an involution.
Archiv der Mathematik, 74:241-245, (2000).
[23] J. Brinkman, W. Bosma, J. Cannon and C. Playoust. The magma algebra system.
I. The user language. J. Symbolic Comput. 24(3-4):235-265, (1997).
[24] J. Brinkman. Generators of order two for Sn and its two double covers. Comm.
Algebra. No 1, 41:223-231, (2000).
[25] J. Brinkman. Involutions in the decorations of M12 . Comm. Algebra. 23,(5):19751988, (1995).
[26] D. Bundy. The connectivity of commuting graphs. J. Combin. Theory Ser.
A,113(6):995-1007, (2006).
[27] R. W. Carter. Simple groups of Lie type. Wiley Classics Library. John Wiley and
Sons, Inc., New York, (1989) Reprint of the 1972 original.
[28] R. W. Carter. Finite groups of Lie type. Conjugacy classes and complex characters. Reprint of the 1985 original. Wiley Classics Library. A Wiley-Interscience
Publication. John Wiley and Sons, Ltd., Chichester:pp xii+544 (1993).
[29] T. Tamizh Chelvama, K. Selvakumar and S. Raja. Commuting Graphs on Dihedral
Group. The Journal of Computer Science, V2, N2:402-406, (2011).
[30] G. Chen. On Thompson’s conjecture. J. Algebra, 185(1):184-193, (1996).
[31] H. Conway, R. T. Curtis, S. P .Norton and R. A. Parker. ATLAS of Finite Groups:
Maximal Subgroups and Ordinary Characters for Simple Groups. Clarendon press.
Oxford, (1985).
[32] Computational Algebra Group, School of Mathematics and Statistics, University
of Sydney. The Magma Handbook. http://magma.maths.usyd.edu.au/magma/hand
book/.
Bibliography
165
[33] R. T. Curtis. A new combinatorial approach to M24., Math. Proc. Cambridge
Philos. Soc. 79,no. 1, 2542, (1976).
[34] M. R. Darafsheh, H. Bigdely, A. Bahrami and M. Davoudi Monfared. Some Results
On Non-Commuting Graph Of A Finite Group. Italian Journal of Pure and Applied
Mathematics, n.27:107-118, (2010).
[35] M. R. Darafsheh. Groups with the same non-commuting graph. Discrete
Appl.Math., 157(4):833-837, (2009).
[36] A. Everett. Commuting Involution Graphs of Certain Finite Simple Classical
Groups. PhD thesis, Manchester Institute for Mathematical Sciences, The University
of Manchester, (2011).
[37] A. Everett. Commuting Involution Graphs for 3-Dimensional Unitary Groups. The
Electronic Journal of Combinatorics), 18:P103, (2011).
[38] A. Everett, P.Rowley. Commuting Involution Graphs for 4-dimensional Projective
Symplectic Groups. Manchester Institute for Mathematical Sciences, The University
of Manchester. MIMS EPrint, (2010).
[39] B. Fischer. Finite groups generated by 3-transpositions. Warwick Mathematical
Institute Preprint, (1970).
[40] B. Fischer. Finite groups generated by 3-transpositions. I. Invent. Math.,13:232246, (1971).
[41] The GAP Group. GAP Groups, Algorithms, and Programming,. Version 4.4.12,
(2008). http://www.gap-system.org.
[42] D. J. H. Garling Clifford Algebra: An Introduction. London Mathematical Society
Student Texts (78). Cambridge University Press, Cambridge, Jun (2011).
[43] M. Geck. An introduction to algebraic geometry and algebraic groups. Oxford Graduate Texts in Mathematics, 10. Oxford University Press, Oxford, (2003).
Bibliography
[44] D. Gorenstein.
166
Finite Groups. Second edition, Chelsea Publishing Company,
(1980).
[45] D. F. Holt, B. Eick, and E. A. O′ Brien.Handbook of computational group theory.
Discrete Mathematics and its Applications (Boca Raton). Chapman & Hall/CRC,
Boca Raton, FL, (2005).
[46] J. F. Humphreys Conjugacy classes of double covers of monomial groups. Math.
Proc. Camb. Phil. Soc., 96:195-201, (1984).
[47] A. Iranmanesh, and A. Jafarzadeh. On the commuting graph associated with the
symmetric and alternating groups. J. Algebra Appl., 7(1):129-146, (2008).
[48] D. L. Johnson. Presentations of Groups. Volume 15 of London Mathematical Student Texts. Cambridge University Press, Cambridge, (1990).
[49] F. Lubeck. Conjugacy classes and character degrees of E7(2)., http://www.math.
rwth-aachen.de/Frank.Luebeck/chev/E72.html.
[50] E. L. Mathieu. Memoire sur l′ tude des fonctions de plusieurs quantits, sur la
manire de les former et sur les substitutions qui les laissent invariables. J. Math.
Pures Appl. (Liouville) (2) VI: pp 241-323, (1861).
[51] E. L. Mathieu. Sur la fonction cinq fois transitive de 24 quantits. Liouville Journ.,
(2)XVIII: pp 25-47, (1873).
[52] A. R. Mohaddamfar, W. J. Shi, W. Zhou and A. R. Zokayi. On the non-commuting
graph associated with a finite group. Siberian mathematical journal,V46, NO2:325332, (2005).
[53] A. Nawawi and P. Rowley. On commuting graphs for elements of order 3 in symmetric groups. PhD thesis,University of Manchester, ISSN 1749-9097, (2012).
[54] B. H. Neumann. A problem of Paul Erdos on groups. J. Austral. Math. Soc.Ser.
A, 21(4):467-472, (1976).
Bibliography
167
[55] I. M. Isaacs. Character theory of finite groups. Simple group. Academic Press, New
York, (1976).
[56] C. Parker. The Commuting Graph Of A Soluble Group. Bull. London Math. Soc.
45 (4): 839-848, (2013).
[57] S. Perkins. Commuting involution graphs for Aˆn . Arch. Math. (Basel), 86(1):16-25,
(2006).
[58] C. Procesi. Lie Groups: An Approach Through Invariants and Representations.
Springer Science Business Media, LLC, New York NY, 1st ed., ISSN 0172-5939,
(2007).
[59] S. Rapinchuk, Y. Segev and M. Seitz. Finite quotients of the multiplicative group of
a finite dimensional division algebra are solvable. J. Amer. Math. Soc., 15(4):929-978,
(2002).
[60] P. Rowley. The diameter of the monster graph.(Manuscript).
[61] I. Schur. Darstellungstheorie der Alternatingsche und Symmetrische Gruppen. J.
Mathematik, No.139:155-225, (1911).
[62] R. Solomon and A. Woldar. Simple non-abelian groups are characterized by their
non-commuting graph. Preprint (2012).
[63] R. Struik. A note on prime-power groups.Canadian Mathematical Bulletin 3: 2730,
ISSN 0008-4395, (1960).
[64] P. Taylor. Computational Investigation into Finite Groups. PhD thesis, University
of Manchester, (2011).
[65] P. Taylor. Involutions in Fischer′ s sporadic groups. preprint, http:// eprints.ma.ma
n.ac.uk/1622, (2011).
[66] J. Tits. Algebraic and abstract simple groups. Ann. of Math. (2) 80: 313329,(1964).
Bibliography
168
[67] J. Tripp, I. Suleiman, S. Rogers R. Parker, S. Norton, S. Nickerson, S. Linton,
J. Bray, A. Wilson and P. Walsh. A world wide web atlas of group representations.,
http://brauer.maths.qmul.ac.uk/Atlas/v3/.
[68] D. B. Wales. Some projective representations of Sn . J. Algebra, 61:37-57, (1979).
[69] L. Wang and W. Shi. A new characterization of A10 by its non-commuting graph.
Comm.Algebra, 36:533-540, (2008).
[70] D. Ward. Topics in Finite Groups: Homology Groups, Pi-product Graphs, Wreath
Products and Cuspidal Characters. PhD thesis,University of Manchester, (2015).
[71] R. Wilson. The Finite Simple Groups. Graduate Texts in Mathematics. Springer,
first edition, December (2009).
Notations and Symbols
Symbols
C( −) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
CC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
F ix(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
Sn+ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Sn− . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
XC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
[− − · · · −] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
∆i (−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Γ(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
Λ(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
Ω . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
C(−, −) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Gx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
Diam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
ω(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
ρ(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
θ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
d(−, −) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
f ix(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
supp(−) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
169
Notations and Symbols
170
αi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Ψ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Υ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
ς . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Li . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
O2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Pi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Qi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
Uαi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
xαi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
Xα . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91