Nothing Special   »   [go: up one dir, main page]

Next Article in Journal
Weighted Pseudo-θ-Almost Periodic Sequence and Finite-Time Guaranteed Cost Control for Discrete-Space and Discrete-Time Stochastic Genetic Regulatory Networks with Time Delays
Next Article in Special Issue
First-Order Differential Subordinations and Their Applications
Previous Article in Journal
On Estimation of Reliability Functions for the Extended Rayleigh Distribution under Progressive First-Failure Censoring Model
Previous Article in Special Issue
An Efficient Class of Estimators in Stratified Random Sampling with an Application to Real Data
You seem to have javascript disabled. Please note that many of the page functionalities won't work as expected without javascript enabled.
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Existence, Uniqueness and the Multi-Stability Results for a W-Hilfer Fractional Differential Equation

by
Safoura Rezaei Aderyani
1,
Reza Saadati
1,*,
Themistocles M. Rassias
2,* and
Hari M. Srivastava
3,4,5,6
1
School of Mathematics, Iran University of Science and Technology, Narmak, Tehran 16844, Iran
2
Department of Mathematics, National Technical University of Athens, Zografou Campus, 15780 Athens, Greece
3
Department of Mathematics and Statistics, Universtity of Victoria, Victoria, BC V8P 5C2, Canada
4
Department of Medical Research, China Medical University Hospital, China Medical University, Taichung 40402, Taiwan
5
Department of Mathematics and Informatics, Azerbaijan University, 71 Jeyhun Hajibeyli Street, AZ1007 Baku, Azerbaijan
6
Center for Converging Humanities, Kyung Hee University, 26 Kyungheedae-ro, Dongdaemun-gu, Seoul 02447, Republic of Korea
*
Authors to whom correspondence should be addressed.
Axioms 2023, 12(7), 681; https://doi.org/10.3390/axioms12070681
Submission received: 21 May 2023 / Revised: 26 June 2023 / Accepted: 5 July 2023 / Published: 11 July 2023
(This article belongs to the Special Issue Mathematical Analysis and Applications IV)

Abstract

:
In this paper, we apply the well-known aggregation mappings on Mittag-Leffler-type functions to investigating new approximation error estimates of a W -Hilfer fractional differential equation, by a different concept of Ulam-type stability in both bounded and unbounded domains.
MSC:
39B62; 46L05; 47B47; 47H10; 46L57

1. Introduction

The main issue we are studying in the present paper is that of aggregation mapping, which refers to the procedure of merging some inputs into a single output. Probably the oldest instance is the concept of the arithmetic mean, which has been applied throughout the history of empirical sciences. Any map, such as the arithmetic mean, that computes a unique output value from a vector of input values, is named an aggregation map [1].
Aggregation maps play a significant role in various technical tasks scholars that are faced with in current times. They are particularly significant in regard to the diverse problems relevant to the fusion of information. Generally, aggregation maps are widely applied in applied mathematics (e.g., statistics, probability, decision mathematics), pure mathematics (e.g., theory of means and averages, functional equations, measure theory), social sciences (e.g., mathematical psychology), computer and engineering sciences (e.g., operations research, engineering design, artificial intelligence, information theory, data fusion, image analysis), economics and finance (e.g., game theory, decision making, voting theory) and many other natural sciences. Thus, an important characteristic of aggregation maps is that they are applied in different fields [2,3].
Here, we apply n-ary aggregation maps on well-known special functions, including Mittag-Leffler-type functions, to define a class of matrix-valued controller, which helps us to present a new concept of Ulam-type stability. The aggregation maps allow us to obtain the best approximation error estimates by a different concept of perturbation stability, depending on the variant special functions that are initially chosen, and to study minimal errors and optimal stability, which enables us to obtain a single optimal solution.
The history of Ulam-type stability commenced in the 19th century. This stability was first presented by Stanisław Marcin Ulam [4], for an additive function, which was investigated by Donald Hyers [5], for a group of homomorphisms given on a Banach space. Thereafter, the stability notion was extended by Themistocles Rassias [6], and was named Ulam–Hyers–Rassias (UHR) stability. In addition, Ger and Alsina [7] investigated the Ulam-type stability of ODEs, by replacing functional equations. In [8], Jung and Algifiary proposed the Ulam stability of nth-order ODEs, by means of the Laplace transform technique. Jung, Rezaei and Rassias [9] studied the Ulam stability of ODEs by the Laplace transform technique. Applying the UHR technique, Baleanu and Wu [10] demonstrated the Mittag-Leffler (ML)-type stability of fractional equations; Baleanu, Wu and Huang [11] demonstrated the ML-type stability of fractional delay difference equations with impulse; Wu [12] proved the ML-type stability of fractional neural networks through the fixed point (FP) theory see also [13].
Here, we present some novel notions concerning the stability of fractional equations in the Mittag-Leffler–Hyers–Ulam sense, by the FP technique, which is the most popular technique for studying the stability of different types of equations. The FP technique was applied for the first time by Baker [14], who used it to get the UHR stability of a functional equation in a single variable. At present, numerous authors follow Radu’s technique [15], and make use of a theorem of Margolis and Diaz.
Here, we study existence, uniqueness and the multi-stability results for the fractional system below:
H D Θ + X , Z ; W J ( S ) = ρ ( S , J ( S ) , H D Θ + X , Z ; W J ( S ) ) , I Θ + 1 𝒲 ; W J ( Θ + ) = J Θ ,
where H D Θ + X , Z ; W ( . ) is a fractional-order derivative in the Hilfer sense of order X ( 0 , 1 ] and type Z [ 0 , 1 ] , I Θ + 1 𝒲 ; W ( . ) is a fractional integral of order 1 𝒲 , 𝒲 = X + Z ( 1 X ) , in regard to the function W defined in Definition 2, and ρ : Υ × R 2 R is an arbitrary function. Presume S Υ = [ Θ , D ] with D > Θ and J Θ R .
Let [ L 1 , L 2 ] ( 0 L 1 < L 2 < ) be an interval, and C [ L 1 , L 2 ] be the space of continuous functions h : [ L 1 , L 2 ] R with norm
h C [ L 1 , L 2 ] = max L 1 ζ L 2 | h ( ζ ) | .
The weighted space C 1 𝒲 ; W [ L 1 , L 2 ] of continuous functions h on ( L 1 , L 2 ] is defined by
C 1 𝒲 ; W [ L 1 , L 2 ] = h : ( L 1 , L 2 ] R ; ( W ( ζ ) W ( L 1 ) ) 1 𝒲 h ( ζ ) C [ L 1 , L 2 ] , 0 𝒲 < 1 ,
with norm
h C 1 𝒲 ; W [ L 1 , L 2 ] = max ζ [ L 1 , L 2 ] | ( W ( ζ ) W ( L 1 ) ) 1 𝒲 h ( ζ ) | .

2. Preliminaries

2.1. On Fractional Derivatives

Definition 1
([1]). Let the interval ( L 1 , L 2 ) ( L i i = 1 , 2 R ) , and X > 0 . Presume W ( ζ ) is a monotonically increasing and positive function on ( L 1 , L 2 ] that has a continuous derivative W ( ζ ) on ( L 1 , L 2 ) . The fractional integral respecting W on [ L 1 , L 2 ] is given by
I L 1 + X ; W h ( ζ ) = 1 Γ ( X ) L 1 S W ( S ) ( W ( ζ ) W ( S ) ) X 1 h ( S ) d S .
Definition 2
([1]). Let X ( α 1 , α ) with α N , and let h , W C α [ L 1 , L 2 ] be two functions, such that W is increasing and W ( ζ ) 0 for any ζ [ L 1 , L 2 ] ; therefore, the W -Hilfer fractional derivative H D L 1 + X , Z ; W ( . ) of order X and type Z [ 0 , 1 ] is defined as follows:
H D L 1 + X , Z ; W h ( ζ ) = I L 1 + Z ( α X ) ; W 1 W ( ζ ) d d ζ α I L 1 + ( 1 Z ) ( α X ) ; W h ( ζ ) .
Theorem 1
([1]). Let h C 1 [ L 1 , L 2 ] , X ( 0 , 1 ) and Z [ 0 , 1 ] . Then,
H D L 1 + X , Z ; W I L 1 + X ; W h ( ζ ) = h ( ζ ) .
Theorem 2
([1]). Let h C 1 [ L 1 , L 2 ] , X ( 0 , 1 ) and Z [ 0 , 1 ] . Then,
I L 1 + X ; W H D L 1 + X , Z ; W h ( ζ ) = h ( ζ ) ( W ( ζ ) W ( L 1 ) ) 𝒲 1 Γ ( 𝒲 ) I L 1 + ( 1 Z ) ( 1 X ) ; W h ( L 1 ) .
Lemma 1
([16]). Let η 1 , η 2 > 0 . If h ( S ) = ( W ( S ) W ( L 1 ) ) η 2 1 . Then, we obtain
I L 1 + η 1 ; W h ( S ) = Γ ( η 2 ) Γ ( η 2 + η 1 ) ( W ( S ) W ( L 1 ) ) η 1 + η 2 1 .

2.2. On the Alternative Theory

Theorem 3
([17]). Let n N . Consider the set Υ with a complete [ 0 , ] n -valued metric χ (see [18]), and consider the self-map ∝ on Υ , satisfying the inequality
χ ( λ 1 , λ 2 ) ( β , , β ) n χ ( λ 2 , λ 1 ) , β < 1 is a Lipschitz constant ,
for every λ 1 , λ 2 Υ . Thus, we have two options:
(I) χ ( m λ 1 , m + 1 λ 1 ) = ( + , , + ) n , m N ,
or
(II) we obtain m 0 N s.t.:
( 1 )   χ ( m λ 1 , m + 1 λ 1 ) < ( + , , + ) n , m m 0 ;
( 2 ) the fixed point λ 2 of ∝ is the convergent point of the sequence { m λ 1 } ;
( 3 )   λ 2 is the unique fixed point of ∝, in the set V = { λ 2 Υ χ ( m 0 λ 1 , λ 2 ) < ( + , , + ) n } ;
( 4 )   ( ( 1 β ) , , ( 1 β ) ) n χ ( λ 2 , λ 2 ) χ ( λ 2 , λ 2 ) , for every λ 2 Υ .

2.3. On Aggregation Maps and Special Functions

Firstly, we introduce the concept of aggregation maps. Next, we apply a small list of aggregation maps, to study optimal stability, which helps us to obtain a unique optimum solution.
Now, let
diag [ Δ 1 , , Δ n ] : = diag Δ 1 Δ n , n N .
Note that diag [ Δ 1 , , Δ n ] diag [ 1 , , n ] , iff Δ i i , for any i = 1 , , n .
Let n N and [ n ] : = { 1 , , n } . We usually apply bold symbols to demonstrate n-tuples: for instance, diag [ y 1 , , y n ] n × n will usually be written Y . Let J be a real interval.
Definition 3
([19]). A function P ( n ) : diag [ J , , J ] n × n J is an aggregation function, if:
(i) 
it is nondecreasing in each variable;
(ii) 
it satisfies the boundary conditions
sup Y J n P ( n ) ( Y ) = sup J , and inf Y J n P ( n ) ( Y ) = inf J .
n N shows the arty of the aggregation map. Note that the aggregation maps will be written P instead of P ( n ) .
We present a common list of aggregation maps, as follows:
  • The geometric mean function GM : diag [ J , , J ] n × n J and the arithmetic mean function AM : diag [ J , , J ] n × n J are, respectively, given by
    AG 1 ( Y ) : = AM ( Y ) : = 1 n i = 1 n y i ,
    AG 2 ( Y ) : = GM ( Y ) : = ( i = 1 n y i ) 1 n ;
  • For every k [ n ] , the projection function P k : diag [ J , , J ] n × n J and the order statistic function OS k : diag [ J , , J ] n × n J related to the k th argument, are correspondingly given by
    AG 3 ( Y ) : = P k ( Y ) : = y k ,
    AG 4 ( Y ) : = OS k ( Y ) : = ( y ) k ,
    where ( y ) k is the k th lowest coordinate of y:
    y ( 1 ) y ( k ) y ( n ) .
    The projections onto the first and the last coordinates are given by
    AG 5 ( Y ) : = P F ( Y ) : = P 1 ( Y ) = y 1 ,
    AG 6 ( Y ) : = P L ( Y ) : = P n ( Y ) = y n .
    Likewise, the extreme order statistics y 1 and y n are correspondingly the minimum and maximum functions
    AG 7 ( Y ) : = Min ( Y ) : = OS 1 ( Y ) = min { y 1 , , y n } ,
    AG 8 ( Y ) : = Max ( Y ) : = OS n ( Y ) = max { y 1 , , y n } ,
    which will be written through the operations ∨ and ∧, respectively:
    Max ( Y ) = i = 1 n y i , and Min ( Y ) = i = 1 n y i .
    Similarly, the median of odd numbers of values diag [ y 1 , , y 2 k 1 ] ( 1 + 2 k ) × ( 1 + 2 k ) is simply given by
    Med ( diag [ y 1 , , y 1 + 2 k ] ( 1 + 2 k ) × ( 2 k 1 ) ) = y ( k ) .
    For an even number of values diag [ y 1 , , y 2 k ] , the median is given by
    Med diag [ y 1 , , y 2 k ] 2 k × 2 k : = AM diag [ y ( k + 1 ) , y ( k ) ] 2 × 2 = y ( k + 1 ) + y ( k ) 2 .
    For every φ J , we also define the φ -median, Med φ : diag [ J , , J ] n × n J , by
    AG 9 ( Y ) : = Med φ ( Y ) = Med diag [ y 1 , , y n , φ , , φ n 1 ] ( 2 n 1 ) × ( 2 n 1 ) = Med ( Min ( Y ) , φ , Max ( Y ) ) ;
  • For every K [ n ] , the partial minimum Min k : diag [ J , , J ] n × n J and the partial maximum Max k : diag [ J , , J ] n × n J , associated with K, are, respectively, given by
    AG 10 ( Y ) : = Min k ( Y ) : = i K y i ,
    AG 11 ( Y ) : = Max k ( Y ) : = i K y i ;
  • For every weight vector V = diag [ v 1 , , v n ] n × n diag [ [ 0 , 1 ] , , [ 0 , 1 ] ] n × n s.t. i = 1 n v i = 1 , the weighted arithmetic mean function
    WAM V : diag [ J , , J ] n × n J
    and the ordered weighted averaging function OWA V : diag [ J , , J ] n × n J , associated with V , are, respectively, given by
    AG 12 ( Y ) : = WAM V ( Y ) : = i = 1 n v i y i , AG 13 ( Y ) : = OWA V ( Y ) : = i = 1 n v i y ( i ) ;
  • The sum and product functions , : diag [ R ¯ , , R ¯ ] n × n R are correspondingly given by
    AG 14 ( Y ) : = ( Y ) : = i = 1 n y i , AG 15 ( Y ) : = Π ( Y ) : = i = 1 n y i .
The main issue we are studying in this section is that of aggregation mapping, which refers to the process of combining various input values into a single output. We will apply the above aggregation mappings on Mittag-Leffler functions, to study the stability results for the governing model.
Consider the following special functions:
  • The one-parameter Mittag-Leffler function [20],
    ϶ 1 ( Λ ) : = X ( Λ ) = i = 0 Λ i Γ ( i X + 1 ) ,
    in which Λ , X C , i N , and ( ) > 0 ;
  • The pre-superhyperbolic supercosine through (17) [20],
    ϶ 2 ( Λ ) : = p r e c o s h X ( Λ ) = 0.5 ( X ( Λ ) + X ( Λ ) ) = i = 0 Λ 2 i Γ ( ( 2 i ) X + 1 ) ,
    where Λ , X C , and ( X ) > 0 ;
  • The pre-supercosine function through (17) [20],
    ϶ 3 ( Λ ) : = p r e c o s X ( Λ ) = 1 2 ( X ( i Λ ) + X ( i Λ ) ) = i = 0 ( 1 ) i Λ 2 i Γ ( ( 2 i ) X + 1 ) ,
    where Λ , X C , and ( X ) > 0 ;
  • The pre-superhyperbolic supersine through (17) [20],
    ϶ 4 ( Λ ) : = p r e s i n h X ( Λ ) = 1 2 ( X ( Λ ) X ( Λ ) ) = i = 0 Λ 2 i + 1 Γ ( ( 2 i + 1 ) X + 1 ) ,
    in which Λ , X C , and ( X ) > 0 ;
  • The pre-supersine function through (17) [20],
    ϶ 5 ( Λ ) : = p r e s i n X ( Λ ) = 1 2 i ( X ( i Λ ) X ( i Λ ) ) = i = 0 ( 1 ) i Λ 2 i + 1 Γ ( ( 2 i + 1 ) X + 1 ) ,
    where Λ , X C , and ( X ) > 0 .
Here, we define the matrix-valued controller S as follows:
S ( Λ ) : = diag [ ϶ 1 ( Λ ) , , ϶ 5 ( Λ ) ] .
Note that we have the following inequalities:
θ Γ ( X ) 0 S W ( s ) ( W ( S ) W ( s ) ) X 1 X ( W ( s ) W ( 0 ) ) X d s θ Γ ( X ) 0 S W ( s ) ( W ( S ) W ( s ) ) X 1 k = 0 ( W ( s ) W ( 0 ) ) k X Γ ( k X + 1 ) d s = θ Γ ( X ) k = 0 1 Γ ( k X + 1 ) 0 S ( W ( S ) W ( s ) ) X 1 ( W ( s ) W ( 0 ) ) k X d W ( s ) = θ Γ ( X ) k = 0 1 Γ ( k X + 1 ) 0 W ( S ) W ( 0 ) ( W ( S ) W ( 0 ) u ) X 1 ( u ) k X d u ( u = W ( s ) W ( 0 ) ) θ Γ ( X ) k = 0 1 Γ ( k X + 1 ) ( W ( S ) W ( 0 ) ) X 1 0 W ( S ) W ( 0 ) ( 1 u W ( S ) W ( 0 ) ) X 1 u k X d u = θ Γ ( X ) k = 0 1 Γ ( k X + 1 ) ( W ( S ) W ( 0 ) ) ( k + 1 ) X 0 1 ( 1 v ) X 1 v k X d v v = u W ( S ) W ( 0 ) = θ Γ ( X ) k = 0 1 Γ ( k X + 1 ) ( W ( S ) W ( 0 ) ) ( k + 1 ) X Γ ( k X + 1 ) Γ ( X ) Γ ( ( k + 1 ) X + 1 ) θ n = 0 ( W ( S ) W ( 0 ) ) n X Γ ( n X + 1 ) = θ X ( W ( S ) W ( 0 ) ) X .
Now, we obtain
θ Γ ( X ) 0 S W ( s ) ( W ( S ) W ( s ) ) X 1 S ( W ( s ) W ( 0 ) ) X d s θ S ( W ( S ) W ( 0 ) ) X .

3. Existence, Uniqueness and Multi-Stability

Making use of Theorem 3, we study existence, uniqueness and the multi-stability results of the system (19) and (20).
Now, for ρ : Υ × R 2 R and 0 < θ , we consider the following equations:
H D Θ + X , Z ; W J ( S ) = ρ ( S , J ( S ) , H D Θ + X , Z ; W J ( S ) ) ;
I Θ + 𝒲 ; W J ( Θ + ) = J Θ , J Θ R ,
and the following inequality:
diag [ | H D Θ + X , Z ; W L ( S ) ρ ( S , L ( S ) , H D Θ + X , Z ; W L ( S ) ) | , , | H D Θ + X , Z ; W L ( S ) ρ ( S , L ( S ) , H D Θ + X , Z ; W L ( S ) ) | ] 15 × 15 diag θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ,
where S Υ .
Now, we define the notion of multi-stability.
Definition 4.
Equations (19) and (20) have multi-stability with respect to
diag AG 1 S ( W ( S ) W ( 0 ) ) X , , AG 15 S ( W ( S ) W ( 0 ) ) X ,
if there exists c i i = 1 , , 15 > 0 , such that for every θ i i = 1 , , 15 > 0 and every solution L C 1 𝒲 ; W ( Υ , R ) to (45) and I Θ + 1 𝒲 ; W L ( Θ + ) = J Θ , there exists a solution J C 1 𝒲 ; W ( Υ , R ) to (19) and (20) with
diag | L ( S ) J ( S ) | , , | L ( S ) J ( S ) | 15 × 15 diag c 1 θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , c 15 θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ,
for every S δ .
Lemma 2
([21]). Let a continuous function ρ ( S , κ , υ ) : Υ × R 2 R . Then, Equation (1) is equivalent to
J ( S ) = ( W ( S ) W ( Θ ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + I Θ + X ; W ( S ) ,
where C ( Υ , R ) satisfies the equality below:
J ( S ) = ρ S , J ( S ) , J ( S ) .
Remark 1.
Let L C 1 𝒲 ; W ( Υ , R ) be a solution of (45), and I Θ + 1 𝒲 ; W L ( Θ + ) = J Θ . Then, L is a solution of the inequality below:
diag [ | L ( S ) ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( s ) ) X 1 L ( s ) d s | , , | L ( S ) ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( s ) ) X 1 L ( s ) d s | ] 15 × 15 diag θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ,
where L C ( Υ , R ) satisfies the equality below:
L ( S ) = ρ S , L ( S ) , L ( S ) .
Let us suppose the following axioms are satisfied:
( F 1 ) ρ : Υ × R 2 R is continuous;
( F 2 ) There is 0 < τ 1 , τ 2 , with 0 < τ 1 ( 1 τ 2 ) < 1 s.t.
| ρ ( S , Ω , ϖ ) ρ ( S , Ω ¯ , ϖ ¯ ) | τ 1 | Ω Ω ¯ | + τ 2 | ϖ ϖ ¯ | for each Ω , ϖ , Ω ¯ , ϖ ¯ R and S Υ .
Theorem 4.
Let ( F 1 ) and ( F 2 ) be satisfied. If L C 1 𝒲 ; W ( Υ , R ) satisfies (45) and I Θ + 1 𝒲 ; W L ( Θ + ) = J Θ , then there is a single function J satisfying (19) and (20), s.t.,
diag | J ( S ) L ( S ) | , , | J ( S ) L ( S ) | 15 × 15 diag [ θ 1 1 τ 1 ( 1 τ 2 ) AG 1 S ( W ( S ) W ( 0 ) ) X , , θ n 1 τ 1 ( 1 τ 2 ) AG 15 S ( W ( S ) W ( 0 ) ) X ] ,
for every S Υ .
Proof. 
Set ϱ = C 1 𝒲 ; W ( Υ , R ) , and define a mapping, χ : ϱ × ϱ [ 0 , ] n , by
χ ( Φ 1 , Φ 2 ) = inf { ( 1 , , 15 ) 0 : diag | Φ 1 ( S ) Φ 2 ( S ) | , , | Φ 1 ( S ) Φ 2 ( S ) | 15 × 15 diag [ 1 θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , 15 θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X } .
We show ( ϱ , χ ) is a complete metric space. Let χ ( Φ 1 , Φ 2 ) > χ ( Φ 1 , ν ) + χ ( ν , Φ 2 ) , for some Φ 1 , Φ 2 and ν ϱ . Thus, there exists S δ with
diag | Φ 1 ( S ) Φ 2 ( S ) | , , | Φ 1 ( S ) Φ 2 ( S ) | 15 × 15 > ( χ ( Φ 1 , ν ) + χ ( ν , Φ 2 ) ) diag [ θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ] .
Thus, from the definition of χ , we obtain
| Φ 1 ( S ) Φ 2 ( S ) | > | Φ 1 ( S ) ν ( S ) | + | ( ν ( S ) Φ 2 ( S ) | ,
which is contradictory. We now show ( ϱ , χ ) is complete. Presume ω k is a Cauchy sequence in ( ϱ , χ ) . Thus, for every ϵ i i = 1 , , 15 > 0 , there exists a ϵ i i = 1 , , 15 N , s.t. χ ( ω m , ω k ) ( ϵ 1 , , ϵ 15 ) , for every m , k ϵ i i = 1 , , 15 . According to (25), we obtain
diag | ω m ( S ) ω k ( S ) | , , | ω m ( S ) ω k ( S ) | 15 × 15 diag ϵ 1 θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , ϵ 15 θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ,
for each S Υ . If S is fixed, (26) concludes that { ω k ( S ) } is a Cauchy sequence in R . As the set of real numbers R is complete, { ω k ( S ) } converges for any S Υ . Then, we obtain a function ω , defined by
ω ( S ) : = lim k ω k ( S ) , ( S Υ ) ,
which gives us ω ϱ , (because { ω k ( S ) } is Cauchy in complete space R , so they are uniformly convergent on the mapping ω defined in (27). The uniform convergence leads us to the fact that ω is continuous, and is an element of ϱ ). If we set m , it follows from (26) that
diag | ω ( S ) ω k ( S ) | , , | ω ( S ) ω k ( S ) | 15 × 15 diag ϵ 1 θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , ϵ 15 θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X .
Considering (25), we obtain
χ ( ω , ω k ) ( ϵ 1 , , ϵ 15 ) .
This confirms that the Cauchy sequence { ω k } converges to ω in ( ϱ , χ ) . Thus, ( ϱ , χ ) is complete. In view of Lemma 2, Equations (19) and (20) are equivalent to the system below:
J ( S ) = ( W ( S ) W ( Θ ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + I Θ + X ; W ( S ) ,
in which C ( Υ , R ) satisfies the equality,
( S ) = ρ S , ( W ( S ) W ( Θ ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + I Θ + X ; W ( S ) , ( S ) ,
for all S Υ . To prove this note, by using I Θ + X ; W ( . ) on both sides of (1), and utilizing Theorem 2, we obtain
J ( S ) ( W ( S ) W ( Θ ) ) 𝒲 1 Γ ( 𝒲 ) I Θ + ( 1 X ) ( 1 Z ) ; W J Θ = I Θ + X ; W ( S ) .
Thus,
J ( S ) = ( W ( S ) W ( Θ ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + I Θ + X ; W ( S ) .
In addition, if J satisfies (30), then it satisfies (1). To see this, apply H D Θ + X , Z ; W ( . ) on both sides of Equation (30). Then, according to Theorem 1, we obtain
H D Θ + X , Z ; W J ( S ) = H D Θ + X , Z ; W ( W ( S ) W ( Θ ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + H D Θ + X , Z ; W I Θ + X ; W ( S ) = ( S ) , where, for 𝒲 ( 0 , 1 ) , we apply H D Θ + X , Z ; W ( W ( S ) W ( Θ ) ) 𝒲 1 = 0 . We deduce that J ( S ) satisfies (1) if J ( S ) satisfies (29).
Let : ϱ ϱ , such that Φ 1 ϱ
( Φ 1 ( S ) ) = ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 Φ 1 ( ϶ ) d ϶ ,
where Φ 1 C ( Υ , R ) satisfies the following equality:
Φ 1 ( S ) = ρ S , Φ 1 ( S ) , Φ 1 ( S ) .
For Φ 1 ϱ , we obtain
| ( Φ 1 ( S ) ) ( Φ 1 ( S 0 ) ) | = | ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 Φ 1 ( ϶ ) d ϶ ( W ( S 0 ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + 1 Γ ( X ) 0 S W ( X ) ( W ( S 0 ) W ( ϶ ) ) X 1 Φ 1 ( ϶ ) d ϶ | 0 , as S S 0 ,
so : ϱ ϱ is continuous.
We now prove ∝ is contractive on ϱ . Let : ϱ ϱ defined in (31). Let Φ 1 , Φ 2 C 1 𝒲 ; W ( Υ , R ) , and χ ( Φ 1 ( S ) , Φ 2 ( S ) ) ( k 1 , , k 15 ) , and k 1 , , k 15 [ 0 , + ] . Then, for all S Υ , we obtain
diag | Φ 1 ( S ) Φ 2 ( S ) | , , | Φ 1 ( S ) Φ 2 ( S ) | 15 × 15 diag k 1 θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , k 15 θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X .
For all S Υ , we obtain
| ( Φ 1 ( S ) ) ( Φ 2 ( S ) ) | 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 | Φ 1 ( ϶ ) Φ 2 ( ϶ ) | d ϶ ,
where
Φ 1 ( S ) = ρ ( S , Φ 1 ( S ) , Φ 1 ( S ) ) ,
and
Φ 2 ( S ) = ρ ( S , Φ 2 ( S ) , Φ 2 ( S ) ) .
Through hypothesis ( F 2 ) , we obtain, for any S Υ ,
| Φ 1 ( S ) Φ 2 ( S ) | τ 1 | Φ 1 ( S ) Φ 2 ( S ) | + τ 2 | Φ 1 ( S ) Φ 2 ( S ) | ,
which can be written as
| Φ 1 ( S ) Φ 2 ( S ) | τ 1 1 τ 2 | Φ 1 ( S ) Φ 2 ( S ) | .
Next, using Remark 1, (33) and (34), we obtain
diag | ( Φ 1 ( S ) ) ( Φ 2 ( S ) ) | , , | ( Φ 1 ( S ) ) ( Φ 2 ( S ) ) | 15 × 15 diag [ τ 1 ( 1 τ 2 ) Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 | Φ 1 ( ϶ ) Φ 2 ( ϶ ) | d ϶ , , τ 1 ( 1 τ 2 ) Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 | Φ 1 ( ϶ ) Φ 2 ( ϶ ) | d ϶ ] 15 × 15 diag [ k 1 θ 1 τ 1 ( 1 τ 2 ) 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 × AG 1 ( S ( W ( ϶ ) W ( 0 ) ) X ) d ϶ , , k 15 θ 15 τ 1 ( 1 τ 2 ) 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 × AG 15 ( S ( W ( ϶ ) W ( 0 ) ) X ) d ϶ ] diag [ k 1 θ 1 τ 1 ( 1 τ 2 ) AG 1 S ( W ( S ) W ( 0 ) ) X , , k 15 θ 15 τ 1 ( 1 τ 2 ) AG 15 S ( W ( S ) W ( 0 ) ) X ] .
Then, we obtain
χ ( ( Φ 1 ) , ( Φ 2 ) ) ( τ 1 ( 1 τ 2 ) , , τ 1 ( 1 τ 2 ) ) 15 χ ( Φ 1 , Φ 2 ) .
As 0 < τ 1 ( 1 τ 2 ) < 1 , we deduce the contractive property of ∝.
Let L ϱ . As ( L ) ϱ , we obtain
diag | ( L ( S ) ) L ( S ) | , , | ( L ( S ) ) L ( S ) | 15 × 15 diag [ | L ( S ) ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 L ( ϶ ) d ϶ | , , | L ( S ) ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 L ( ϶ ) d ϶ | ] 15 × 15 diag θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ,
for S Υ , which implies that
χ ( ( L ) , L ) ( 1 , , 1 ) 15 ;
hence, for all k N , we obtain χ ( k ( L ) , k + 1 ( L ) ) < ( + , , + ) 15 . We now use Theorem 3, and so we obtain a single map J { σ ˜ ϱ : χ ( L , σ ˜ ) < ( + , , + ) 15 } , such that J = J . Thus,
J ( S ) = ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + 1 Γ ( X ) 0 S W ( ϶ ) ( W ( S ) W ( ϶ ) ) X 1 J ( ϶ ) d ϶ ,
for every S Υ , where J C ( Υ , R ) satisfies the equality (32), and I Θ + 1 𝒲 ; W J ( Θ + ) = J Θ R .
Based on Theorem 3 and (35), we obtain
χ ( J , L ) 1 1 τ 1 ( 1 τ 2 ) , , 1 1 τ 1 ( 1 τ 2 ) 15 χ ( ( L ) , L ) ( 1 1 τ 1 ( 1 τ 2 ) , , 1 1 τ 1 ( 1 τ 2 ) 15 ,
which concludes (24). □
We study the next theorem for the set of real numbers. By a similar method, we can investigate the theorem for [ 0 , + ) and ( , 0 ] . Let = C 1 𝒲 ; W ( R ) .
Theorem 5.
Let ( F 1 ) and ( F 2 ) be satisfied. If L in satisfies (45), and
I Θ + 1 𝒲 ; W L ( Θ + ) = J Θ ,
then there exists a single function J satisfying (19) and (20), and s.t. (24) is satisfied for all S R .
Proof. 
For all k N , we consider P k = [ I k , I + k ] . Based on Theorem 4, there exists a single function J k C 1 𝒲 ; W ( P k ) , s.t.,
H D Θ + X , Z ; W J n ( S ) = 𝒲 ( S , J k ( S ) , J k ( η ( S ) ) ) , S P k ,
I Θ + 1 𝒲 ; W J k ( Θ + ) = J Θ , J Θ R ,
and
diag | J k ( S ) L ( S ) | , , | J k ( S ) L ( S ) | 15 × 15 diag [ θ 1 1 τ 1 ( 1 τ 2 ) AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 1 τ 1 ( 1 τ 2 ) AG 15 S ( W ( S ) W ( 0 ) ) X ] ,
for all S P k . The uniqueness of J k implies that if S P k , then
J k ( S ) = J k + 1 ( S ) = J k + 2 ( S ) = .
Consider k ( S ) N as
k ( S ) = min { k N | S P k } .
In addition, consider a function J given by
J ( S ) = J k ( S ) ( S ) , S R ,
and we claim J . For S 1 R , we let the integer k 1 = k ( S 1 ) . Thus, S 1 belongs to the interior of P k 1 + 1 , and there is an ϵ > 0 , such that J ( S ) = J k 1 + 1 ( S ) for all S with S 1 ϵ < S < S 1 + ϵ . Then, we prove that J satisfies (19), (20) and (24) for all S R . For all S R , allow the integer k ( S ) . Thus, S P k ( S ) , and we infer from (37) and (38) that
J ( S ) = J k ( S ) ( S ) = ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 J n ( S ) ( ϶ ) d ϶ = ( W ( S ) W ( 0 ) ) 𝒲 1 Γ ( 𝒲 ) J Θ + 1 Γ ( X ) 0 S W ( X ) ( W ( S ) W ( ϶ ) ) X 1 J ( ϶ ) d ϶ ,
where
J ( S ) = ρ ( S , J ( S ) , J ( S ) )
and
J k ( S ) ( S ) = ρ ( S , J k ( S ) ( S ) , J k ( S ) ( S ) )
are in C ( Υ , R ) . The above, (41), is true because k ( ϶ ) k ( S ) for all ϶ P k ( S ) , and we deduce from (40) that
J ( ϶ ) = J k ( ϶ ) ( ϶ ) = J k ( S ) ( ϶ ) .
As J ( S ) = J k ( S ) ( S ) and S P k ( S ) for all S R , (39) concludes that
diag | J ( S ) L ( S ) | , , | J ( S ) L ( S ) | 15 × 15 = diag | J k ( S ) ( S ) L ( S ) | , , | J k ( S ) ( S ) L ( S ) | 15 × 15 diag [ θ 1 1 τ 1 ( 1 τ 2 ) AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 1 τ 1 ( 1 τ 2 ) AG 15 S ( W ( S ) W ( 0 ) ) X ] .
Eventually, we claim J is single. Let J be another function satisfying (19), (20) and (24), for any S R . As J | P k ( S ) ( = J k ( S ) ) and J | P k ( S ) both satisfy (19), (20) and (24) for any S P k ( S ) , and the uniqueness of J k ( S ) = J | P k ( S ) , we conclude that
J ( S ) = J | P k ( S ) ( S ) = J | P k ( S ) ( S ) = J ( S ) ,
as required. □

4. Example

Example 1.
Consider the system (1) for W ( S ) = S 2 , Θ = 0 , D = 1 , X = Z = 1 2 , J : [ 0 , 1 ] R and ρ : [ 0 , 1 ] × R 2 R defined by
ρ ( S , ϕ , ψ ) = 1 5 2 F 1 ( P , Q , Q ; S ) ϕ + 1 8 ψ ,
and we obtain
H D 0 + 1 2 , 1 2 ; S 2 J ( S ) = 1 5 2 F 1 ( P , Q , Q ; S ) J ( S ) + 1 8 H D 0 + 1 2 , 1 2 ; S 2 J ( S ) , I 0 + 1 4 ; S 2 J ( 0 + ) = J 0 R ,
in which P R + and Q R , and 2 F 1 ( P , Q , Q ; . ) is the hypergeometric function.
In addition, consider the inequality below:
diag [ | H D 0 + 1 2 , 1 2 ; S 2 L ( S ) 1 5 2 F 1 ( P , Q , Q ; S ) L ( S ) 1 8 H D 0 + 1 2 , 1 2 ; S 2 L ( S ) | , , | H D 0 + 1 2 , 1 2 ; S 2 L ( S ) 1 5 2 F 1 ( P , Q , Q ; S ) L ( S ) 1 8 H D 0 + 1 2 , 1 2 ; S 2 L ( S ) | ] 15 × 15 diag θ 1 AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 AG 15 S ( W ( S ) W ( 0 ) ) X ,
for every S [ 0 , 1 ] .
For any κ , υ , κ ¯ , υ ¯ R and S [ 0 , 1 ] , we obtain
| ρ ( S , Ω , ϖ ) ρ ( S , Ω ¯ , ϖ ¯ ) | = | 1 5 2 F 1 ( P , Q , Q ; S ) Ω + 1 8 H D 0 + 1 2 , 1 2 ; S 2 ϖ 1 5 2 F 1 ( P , Q , Q ; S ) Ω ¯ 1 8 H D 0 + 1 2 , 1 2 ; S 2 ϖ ¯ | 1 5 2 F 1 ( P , Q , Q ; S ) | Ω Ω ¯ | + 1 8 | ϖ ϖ ¯ | .
Thus, condition ( F 2 ) is satisfied with τ 1 = 1 5 2 F 1 ( P , Q , Q ; S ) and τ 2 = 1 8 , and the condition
8 35 2 F 1 ( P , Q , Q ; S ) < 1
is satisfied. Theorem 4 implies that (44) has a single solution, and is stable with
diag | J ( S ) L ( S ) | , , | J ( S ) L ( S ) | 15 × 15 diag [ θ 1 1 8 35 2 F 1 ( P , Q , Q ; S ) AG 1 S ( W ( S ) W ( 0 ) ) X , , θ 15 1 8 35 2 F 1 ( P , Q , Q ; S ) AG 15 S ( W ( S ) W ( 0 ) ) X ] 15 × 15 ,
where S [ 0 , 1 ] .

5. Conclusions

We used the aggregation maps on diverse special functions such as the Mittag-Leffler function, supertrigonometric and superhyperbolic functions, to propose a novel controller that helps us study a different notion of stability: namely, multi-stability. Multi-stability enables us to obtain various approximations, depending on various special functions, and to obtain optimal stability, which, in turn, enables us to obtain a unique optimum solution.

Author Contributions

All of the authors conceived the study, participated in its design and coordination, drafted the manuscript and participated in the sequence alignment. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Aderyani, S.R.; Saadati, R. Stability and controllability results by n-ary aggregation functions in matrix valued fuzzy n-normed spaces. Inf. Sci. 2023, 643, 119265. [Google Scholar] [CrossRef]
  2. Kurač, Z. Transfer-stable aggregation functions: Applications, challenges, and emerging trends. Decis. Anal. J. 2023, 7, 100210. [Google Scholar] [CrossRef]
  3. Halaš, R.; Mesiar, R.; Pócs, J. On the number of aggregation functions on finite chains as a generalization of Dedekind numbers. Fuzzy Sets Syst. 2023, 466, 108441. [Google Scholar] [CrossRef]
  4. Ulam, S.M. Problem in Modern Mathematics; Science Editors; Willey: New York, NY, USA, 1960. [Google Scholar]
  5. Hyers, D.H. On the stability of linear functional equation. Proc. Natl. Acad. Sci. USA 1924, 27, 222–224. [Google Scholar] [CrossRef] [Green Version]
  6. Rassia, T.M. On the stability of the linear mapping in Banach spaces. Proc. Am. Math. Soc. 1978, 72, 297–300. [Google Scholar] [CrossRef]
  7. Alsina, C.; Ger, R. On some inequalities and stability results related to the exponential function. J. Inequal. Appl. 1998, 2, 373–380. [Google Scholar] [CrossRef]
  8. Algifiary, Q.H.; Jung, S. Laplace transform and generalized Hyers-Ulam stability of linear differential equation. Electron. J. Differ. Equ. 2014, 2014, 1–11. [Google Scholar]
  9. Razaei, H.; Jung, S.; Rassias, T.M. Laplace transform and Hyers-Ulam stability of linear differential equation. J. Math. Anal. Appl. 2013, 403, 244–251. [Google Scholar] [CrossRef]
  10. Wu, G.; Baleanu, D. Stability analysis of impulsive fractional difference equations. Fract. Calc. Appl. Anal. 2018, 21, 354–375. [Google Scholar] [CrossRef]
  11. Wu, G.; Baleanu, D.; Huang, L. Novel Mittag-Leffler stability of linear fractional delay difference equations with impluse. Appl. Math. Lett. 2018, 82, 71–78. [Google Scholar] [CrossRef]
  12. Wu, G.; Abdeljawad, T.; Liu, J.; Baleanu, D.; Wu, K. Mittag-Leffler stability analysis of fractional discrete-time neural networks via fixed point technique. Nonlinear Anal. Model. Control 2019, 24, 919–936. [Google Scholar] [CrossRef]
  13. Srivastava, H.M.; da Costa Sousa, J.V. Multiplicity of Solutions for Fractional-Order Differential Equations via the κ(x)-Laplacian Operator and the Genus Theory. Fractal Fract. 2022, 6, 481. [Google Scholar] [CrossRef]
  14. Mihet, D. On the stability of the additive Cauchy functional equation in random normed spaces. Appl. Math. Lett. 2011, 24, 2005–2009. [Google Scholar] [CrossRef] [Green Version]
  15. Park, C. Stability of some set-valued functional equations. Appl. Math. Lett. 2011, 24, 1910–1914. [Google Scholar] [CrossRef] [Green Version]
  16. Kilbas, A.A.; Srivastava, H.M.; Trujillo, J.J. Theory and Applications of Fractional Differential Equations; Elsevier: Amsterdam, The Netherlands, 2006. [Google Scholar]
  17. Diaz, J.B.; Margolis, B. A fixed point theorem of the alternative, for contractions on a generalized complete metric space. Bull Amer Math Soc. 1968, 74, 305–309. [Google Scholar] [CrossRef] [Green Version]
  18. Rezaei Aderyani, S.; Saadati, R.; Li, C.; Rassias, T.M.; Park, C. Special functions and multi-stability of the Jensen type random operator equation in C*-algebras via fixed point. J. Inequalities Appl. 2023, 2023, 35. [Google Scholar] [CrossRef]
  19. Grabisch, M.; Marichal, J.L.; Mesiar, R.; Pap, E. Aggregation Functions; Cambridge University Press: Cambridge, MA, USA, 2009; Volume 127. [Google Scholar]
  20. Harikrishnan, S.; Elsayed, E.M.; Kanagarajan, K.; Vivek, D. A study of Hilfer-Katugampola type pantograph equations with complex order. Ex. Counterexamples 2022, 2, 100045. [Google Scholar] [CrossRef]
  21. Li, C.; Nonlaopon, K.; Hrytsenko, A.; Beaudin, J. On the analytic and approximate solutions for the fractional nonlinear Schrödinger equations. J. Nonlinear Sci. Appl. 2023, 16, 51–59. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Aderyani, S.R.; Saadati, R.; Rassias, T.M.; Srivastava, H.M. Existence, Uniqueness and the Multi-Stability Results for a W-Hilfer Fractional Differential Equation. Axioms 2023, 12, 681. https://doi.org/10.3390/axioms12070681

AMA Style

Aderyani SR, Saadati R, Rassias TM, Srivastava HM. Existence, Uniqueness and the Multi-Stability Results for a W-Hilfer Fractional Differential Equation. Axioms. 2023; 12(7):681. https://doi.org/10.3390/axioms12070681

Chicago/Turabian Style

Aderyani, Safoura Rezaei, Reza Saadati, Themistocles M. Rassias, and Hari M. Srivastava. 2023. "Existence, Uniqueness and the Multi-Stability Results for a W-Hilfer Fractional Differential Equation" Axioms 12, no. 7: 681. https://doi.org/10.3390/axioms12070681

APA Style

Aderyani, S. R., Saadati, R., Rassias, T. M., & Srivastava, H. M. (2023). Existence, Uniqueness and the Multi-Stability Results for a W-Hilfer Fractional Differential Equation. Axioms, 12(7), 681. https://doi.org/10.3390/axioms12070681

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop