Bandwidth Allocation Policy Using The Game Theory Model in Heterogeneous Wireless Networks
Bandwidth Allocation Policy Using The Game Theory Model in Heterogeneous Wireless Networks
Bandwidth Allocation Policy Using The Game Theory Model in Heterogeneous Wireless Networks
ABSTRACT
In Heterogeneous Wireless Network (WHN), a Mobile Device (MD) can be connected with multiple access technologies
available in a particular geographical area. Selection of appropriate access technology is a challenge for the MD where game
theory policy can be applied. Shapley Value is one of the game-theoretic models that can be applied to allocate appropriate
bandwidth for i number of users for k number of services from N access technologies. It proposes the fairest allocation of the
collectively gained profits among the collaborative players in the game. The primary focus is to find the relative contribution of
every player in a cooperative game. This paper has drawn a numerical analysis of Shapley Value for three wireless access
technologies, namely WiFi, Cellular and WiMAX. It has been shown how much bandwidth can be allocated from an access
technology for a required bandwidth for a specific service. It can be noted from the results that 225Kbps, 141.67Kbps and
233.33Kbps can be allocated from WiFi, Cellular and WiMAX respectively to transmit the 600Kbps data in an HWN network.
Keywords:about four key words separated by commas
1. INTRODUCTION
Game theory is a mathematical tool used in designing and modeling of complex scenarios that involves the interaction
of rational decision makers with mutual and possibly conflicting interests [1]. It was initially invented to solve the
complex issue of economic behavior. It has been popular in other fields, including politics, philosophy, military,
sociology and telecommunication due to the effectiveness of its studying complex dynamics among players. Recently, a
large number of issues of wireless communications and networking, particularly security [2]-[4] have been addressed
using game theory and its solution.Wireless resources are insufficient in terms of bandwidth, power and capacity. On
the other hand, increasing number of wireless access terminals, resource scarcity makes competition among mobile
users for required wireless resources very severe. In this context, game theory can significantly better understand and
optimize allocations of resources among the players. In recent years, game theory has been investigated in order to
address wireless communication issues, including power control, resource allocation, MIMO systems, medium access
control, routing, load control, etc. [1]. A classification has been performed based on OSI layer (Physical, Data link,
Network, and Transport) in the light of game theoretic approaches [5]. A detailed discussion has been covered in a
recent book on the broad area of wireless communications and networking domains, including sensor networking,
vehicular networking, power control system, and radio resource management [6].
Heterogeneous wireless communication networks are dynamic in nature in terms of network load, availability,
energy conservation and monetary cost. Both operators and users seek to maximize their payoffs. In HWN
environment, the MD has the options to select the best suitable access point (AP) for its needs based on its preferences.
Considering the multiple scenarios in HWN including the type of users, technology, service provider and applications,
require the development of the new dimension that offers dynamic automatic networks selection [1]. Many solutions
have been proposed to address this multi-criteria decision-making algorithms in HWN. Game theory [7] also can be
used to deal with the complex decision making between the mobile users and the networks for resource allocations in
HWN [8] environment. The payoffs can be estimated using utility functions based on the several decision criteria from
both sides where game theory can be well suited.
The geographical area of a Sn can be denoted as G Sn {1,2,3,4.... G Sn} . Each BS/AP S S n has a transmission
capacity Cn Mbps. There are M numbers of Multimode Devices (MDs) in that particular geographical area denoted as
the set M, M {1, 2,3, 4.......M } . Each MD, m M can get services from its home network and also from other
available networks using multi-homing services. An MD using multi-homing capability can receive its required
bandwidth from all available wireless technologies in that particular geographical area. The bandwidth allocated to MD
m from network n through BS/AP s can be denoted as B mns where m M , n N and s S n .
The network supports the Constant Bit Rate (CBR) and Variable Bit Rate (VBR) services. The CBR connection of
MD m needs a constant bandwidth Bm from all available wireless BSs/APs in that particular geographical area. On the
max
other hand, a VBR call of MD m requires a bandwidth allocation within a maximum value B m and a minimum value
min max
B m . The maximum level of bandwidth B m can be allocated for VBR when sufficient resources are available in that
particular area. However, if the other MDs reach their capacity limitations, the level of bandwidth can be reduced to
min
minimum bandwidth B m to increase the efficient utilization of the networks N.
For each network n BS/AP s in the geographical area, the allocated resources should be such that the total load in its
coverage area is within the network BS/AP capacity limitation Cn, that is given in (1).
Bmns C n
mM s S n , n (1)
For a CBR service, the allocated bandwidth from available BSs/APs should satisfy the application required
bandwidth Bm,
N sn
B mns B m
n 1 s 1 , m M 1 (2)
For VBR service, the allocated bandwidth from the available BSs/APs to an MD should satisfy in between the
min max
minimum B m and maximum B m bandwidth allocation.
N s n
min max
B m B mns B m
n 1 s 1 , m M 2 (3)
It can be said that if the requirements do not meet the equations (1) to (3) for the respective circumstances, the
bandwidth will not be allocated to any particular link.
Any aforementioned constraints must need to be fulfilled for the respective required allocation. The concern arises, how
much bandwidth can be allocated from each network to allocate the required bandwidth for the specific services. Hence,
a game theory model can be used for the allocation of bandwidth in a heterogeneous wireless network. The theoretic
model of game theory and its numerical analysis have been discussed in the following sections.
available bandwidth bi
k , req
av r
Bi , N and the shaping parameter to ensure that the network does not offer too much
bandwidth to the new connection.
When a user cannot be served by a single network, a model has been proposed that enables the user to split its
application traffic between the coalition members using a cooperative Stackelberg game [12]. It is the characteristic
function which is used to express the payoff of the coalition, and the core concept is used in order to analyze the
stability of the allocation.
Shapley Value [13] is one of the fairest models to solve the Nik person bankruptcy game due to the less computational
complexity [11]. It proposes the most equitable allocation of the collectively gained profits among the collaborative
players in the game. The primary focus is to find the relative contribution of every player.
number of players,
vS indicates the coalition utility including player i, and
vS
i refers to the coalition utility
k
Bi The estate (money) of the company Requested bandwidth for kth types of services from ith network
k , min
dj The minimum claim of the jth creditor The minimum bandwidth demand of the jth user
k
dj The actual money paid for the jth user Bandwidth allocation to the new connection in the network i.
5. NUMERICAL ANALYSIS
Game theoretic model, particularly Shapley Value is one of the suitable approaches for allocating bandwidth in HWN.
It allocates the appropriate contribution from an individual network to mobile devices. This research has applied
Shapley Value algorithm and drawn a numerical analysis in HWN namely, WiFi, Cellular and WiMAX.The data
transmission rate for WiFi is 11 Mbps, for cellular CDMA is 2 Mbps and for WiMAX is 50 Mbps in a single cell. It has
been assumed that these three technologies offer a bandwidth for the new connection
k k k
b wifi
250, b cel
150, b
wi max
300
Kbps for all types of k services. A new connection requests for bandwidth
600 Kbps. Based on the characteristic function of (11), the contribution of all coalitions can be calculated as follows.
v 0
v S 600 (13)
It can be noted from the equation 13 that when there is no coalition in characteristic function v(0), the bandwidth
allocation is zero, whereas for all possible coalition in characteristic function v(S), the bandwidth allocation is to 600.
This numerical analysis consists of three wireless networks and have formed a coalition. The contribution of bandwidth
allocation of three individual networks to the required networks has been calculated in equation 14. The WiFi network
contributed 150, Cellular 50 and WiMAX 200.
vwifi max 0,600 (300 150) 150
vcel max0,600 (250 300) 50
vwi max max0,600 (250 150) 200
(14)
On the other hand, two networks have formed another set of coalition as depicted in equation 15 and their contribution
has been calculated according to the formula.
vwifi, cel max 0,600 300 300
vwifi, wi max max0,600 150 450
vcel, wi max max0,600 250 350
(15)
All the contributed values have been used in table II to calculate the final contribution of the determination of 600Kbps,
the required bandwidth. It can be said from the table II that the WiFi has contributed 222.25Kbps, Cellular 141.87Kbps
and WIMAX 232.60Kbps according to the calculation.
Table 2: Numerical values using Shapley Formula in HWN
i v |s| N vS vS
i | s| 1!Nki | s|!vS vS i
k
Ni !
i=3 3 3 150 50
i=3 3 3 50 16.66
i=3 3 3 100 33
It can be noted from Table 2 that 225Kbps, 141.67Kbps and 233.33Kbps can be shared from WiFi, Cellular and
WiMAX respectively to transmit the 600Kbps data in an HWN network. Similarly, numerical analysis has been
conducted for different requested bandwidth like 800, 1000, 1200 and 1400 Kbps. The allocated bandwidth versus
request has been shown in Figure 2.
Another assumption has been considered that the video service bandwidth offered by one specific network technology is
600 kbps in a particular geographical area. Six users are employing that service and their maximum request 120, 90,
150, 110, 100, 130 kbps summed up to 700 kbps. Based on the Shapley Value formula, the allocated bandwidth has
been calculated for each user shown in Figure 3. If the users increase or decrease in that particular area and the
specific time, the allocation may be varied due to the congestion of the network. It can be noted that by applying game
theory model in a wireless environment, relatively fair bandwidth allocation can be achieved.
6. CONCLUSION
Game theoretic model, mainly Shapley Value is one of the suitable approaches for allocating bandwidth in HWN. It
allocates the appropriate contribution from an individual network to the mobile devices. This paper has applied Shapley
Value algorithm and drawn a numerical analysis in HWN namely, WiFi, Cellular and WiMAX. It has been shown that
to transmit a required data 600Kbps in an HWN network, 225Kbps, 141.67Kbps and 233.33Kbps can be allocated from
WiFi, Cellular and WiMAX respectively. It proposes the fairest allocation of the collectively gained profits among the
collaborative players in the game.
ACKNOWLEDGEMENT
This work was partially supported by Ministry of Higher Education Malaysia (Kementerian Pendidikan Tinggi) under
Fundamental Research Grant Scheme (FRGS) number FRGS13-081-0322
References
[1] R. Trestian, O. Ormond, G.M. Muntean, “Game Theory-Based Network Selection: Solutions and Challenges,”
IEEE Communications Surveys & Tutorials,XIV (4),pp. 1212-1231, 2012.
[2] B.U.I. Khan, R.F. Olanrewaju, M.U. Mattoo, A.A. Aziz, S.A. Lone, “Modeling Malicious Multi-Attacker Node
Collusion in Manets via Game Theory,” Middle-East Journal of Scientific Research, XXV (3), pp. 568-579,
2017.
[3] B.U.I. Khan, R.F. Olanrewaju, F. Anwar, A. Shah, “Manifestation and Mitigation of Node Misbehaviour in Adhoc
Networks,” Wulfenia Journal, XXI (3), pp. 462-470, 2014.
[4] B.U.I. Khan, R.F. Olanrewaju, R.N. Mir, A. Baba, B.W. Adebayo, “Strategic Profiling for Behaviour
Visualization of Malicious Node in Manets using Game Theory,” Journal of Theoretical & Applied Information
Technology, LXXVII (1),pp. 25-43, 2015.
[5] D.E. Charilas, A.D. Panagopoulos, “A Survey on Game Theory Applications in Wireless Networks,” Elsevier
Computer Networks, LIV(18), pp. 3421-3430, 2010.
[6] Y. Zhang, M. Guizani, Game Theory for Wireless Communications and Networking, 1st ed., CRC Press, 2011.
[7] J.V. Neumann, O. Morgenstern, Theory of Games and Economic Behavior, 2nd ed., Princeton, NJ: Princeton
University Press, 1947.
[8] C.P. Maheswaran, C.H. Sulochana, “Utilizing EEM Approach to Tackle Bandwidth Allocation with respect to
Heterogeneous Wireless Networks,” ICT Express, II(2), pp. 80-86, 2016.
[9] M. Pulido, J.S. Soriano, N. Llorca, “Game Theory Techniques for University Management: An Extended
Bankruptcy Model,” Annals of Operations Research, CIX (1-4),pp. 129-142, 2002.
[10] R. Chai, X. Wang, Q.B. Chen, T. Svensson, “Utility-Based Bandwidth Allocation Algorithm for Heterogeneous
Wireless Networks,” Science China Information Sciences, LVI (2), pp. 1-3, 2013.
[11] D. Niyato, E. Hossain, “A Cooperative Game Framework for Bandwidth Allocation in 4G Heterogeneous
Wireless Networks,” In Proceedings of the IEEE Conference on Communications, Istanbul, pp. 4357-4362,
2006.
[12] I.M. Suliman, C. Pomalaza-Raez, I. Oppermann, J. Lehtomaki, “Radio resource allocation in heterogeneous
wireless networks using cooperative games,” Proceedings of the Nordic Radio Symposium on Including Finnish
Wireless Communications Workshop, (NRS/FWCW), 2004.
[13] L.S. Shapley, “A Value for N-Person Games,” in Annals of Mathematics Studies 28, H.W. Kuhn and A.W.
Tucker (eds.), Princeton: Princeton University Press, pp. 307-317, 1953.