
The capital market in Saudi Arabia is fast growing. Assurance of an informed decision while investing in the Saudi Stock Exchange is critical. There has also been an increased quest for advanced decision-making tools due to complexities in selecting a given portfolio, which remains a critical issue of concern among investors in the face of modern investment environment challenges. The research paper offered shall deliver an innovative MCDM technique through which an MCDM model shall be developed in the Saudi Stock Exchange. This MCDM model uses BTIFS with an OWA operator. A novelty of the proposed study is identifying the optimal weight that will be obtained through a newly developed optimization technique known as TFOA. TFOA is a hybrid methodology that brings on board the strengths of DMOA, MPA, and EO for a more precise and efficient calculation of the ideal weights in the portfolio selection process. This would improve the adaptability and effectiveness of the suggested MCDM structure. The effectiveness of the approach is established by comparative analysis with the already existing methods of MCDM, which proves it superior for the optimization of investment portfolios. Sensitivity analysis also conducted to evaluate the strength and dependability of the suggested method. The ranking of weighted portfolios by the ELECTRE method is also, which more establishes the applicability of BTIFS-OWA in real life. The results indicate that the BTIFS-OWA approach along with the TFOA for determining optimal weights provides significant improvements in decision-making accuracy and portfolio optimization compared to traditional methods.
Citation: Sunil Kumar Sharma. A Bhattacharyya Triangular intuitionistic fuzzy sets with a Owa operator-based decision making for optimal portfolio selection in Saudi exchange[J]. AIMS Mathematics, 2024, 9(10): 27247-27271. doi: 10.3934/math.20241324
[1] | D. Ajay, P. Chellamani, G. Rajchakit, N. Boonsatit, P. Hammachukiattikul . Regularity of Pythagorean neutrosophic graphs with an illustration in MCDM. AIMS Mathematics, 2022, 7(5): 9424-9442. doi: 10.3934/math.2022523 |
[2] | Supriya Bhunia, Ganesh Ghorai, Qin Xin . On the fuzzification of Lagrange's theorem in (α,β)-Pythagorean fuzzy environment. AIMS Mathematics, 2021, 6(9): 9290-9308. doi: 10.3934/math.2021540 |
[3] | Akarachai Satirad, Ronnason Chinram, Aiyared Iampan . Pythagorean fuzzy sets in UP-algebras and approximations. AIMS Mathematics, 2021, 6(6): 6002-6032. doi: 10.3934/math.2021354 |
[4] | Supriya Bhunia, Ganesh Ghorai, Qin Xin . On the characterization of Pythagorean fuzzy subgroups. AIMS Mathematics, 2021, 6(1): 962-978. doi: 10.3934/math.2021058 |
[5] | Mehwish Shehzadi, Hanan Alolaiyan, Umer Shuaib, Abdul Razaq, Qin Xin . Crafting optimal cardiovascular treatment strategy in Pythagorean fuzzy dynamic settings. AIMS Mathematics, 2024, 9(11): 31495-31531. doi: 10.3934/math.20241516 |
[6] | Muhammad Zeeshan Hanif, Naveed Yaqoob, Muhammad Riaz, Muhammad Aslam . Linear Diophantine fuzzy graphs with new decision-making approach. AIMS Mathematics, 2022, 7(8): 14532-14556. doi: 10.3934/math.2022801 |
[7] | Abdul Razaq, Ghaliah Alhamzi . On Pythagorean fuzzy ideals of a classical ring. AIMS Mathematics, 2023, 8(2): 4280-4303. doi: 10.3934/math.2023213 |
[8] | Damchaa Adiyanyam, Enkhbayar Azjargal, Lkhagva Buyantogtokh . Bond incident degree indices of stepwise irregular graphs. AIMS Mathematics, 2022, 7(5): 8685-8700. doi: 10.3934/math.2022485 |
[9] | Hifza, Muhammad Gulistan, Zahid Khan, Mohammed M. Al-Shamiri, Muhammad Azhar, Asad Ali, Joseph David Madasi . A new fuzzy decision support system approach; analysis and applications. AIMS Mathematics, 2022, 7(8): 14785-14825. doi: 10.3934/math.2022812 |
[10] | Saleem Abdullah, Muhammad Qiyas, Muhammad Naeem, Mamona, Yi Liu . Pythagorean Cubic fuzzy Hamacher aggregation operators and their application in green supply selection problem. AIMS Mathematics, 2022, 7(3): 4735-4766. doi: 10.3934/math.2022263 |
The capital market in Saudi Arabia is fast growing. Assurance of an informed decision while investing in the Saudi Stock Exchange is critical. There has also been an increased quest for advanced decision-making tools due to complexities in selecting a given portfolio, which remains a critical issue of concern among investors in the face of modern investment environment challenges. The research paper offered shall deliver an innovative MCDM technique through which an MCDM model shall be developed in the Saudi Stock Exchange. This MCDM model uses BTIFS with an OWA operator. A novelty of the proposed study is identifying the optimal weight that will be obtained through a newly developed optimization technique known as TFOA. TFOA is a hybrid methodology that brings on board the strengths of DMOA, MPA, and EO for a more precise and efficient calculation of the ideal weights in the portfolio selection process. This would improve the adaptability and effectiveness of the suggested MCDM structure. The effectiveness of the approach is established by comparative analysis with the already existing methods of MCDM, which proves it superior for the optimization of investment portfolios. Sensitivity analysis also conducted to evaluate the strength and dependability of the suggested method. The ranking of weighted portfolios by the ELECTRE method is also, which more establishes the applicability of BTIFS-OWA in real life. The results indicate that the BTIFS-OWA approach along with the TFOA for determining optimal weights provides significant improvements in decision-making accuracy and portfolio optimization compared to traditional methods.
Fractional calculus has garnered significant attention across various fields of study because of its ability to capture non-local and long-range dependencies in data analysis and modeling. Fractional calculus provides a more flexible framework for capturing intricate dynamics in real-world systems. By incorporating non-integer orders, it becomes possible to model processes that exhibit fractal or self-similar behavior, which are prevalent in many natural and man-made systems [2,18]. This flexibility allows for a more accurate representation of complex phenomena, leading to improved understanding and enhanced predictive capabilities. Moreover, the application of fractional calculus can unveil hidden patterns and long-range dependencies that may go unnoticed when using integer-order operators [3,4,12,13]. In physics, fractional derivatives have been used to model anomalous diffusion [41], viscoelastic materials [17], and complex transport phenomena [20,43].
A wide range of real-world processes exhibit memory effects or non-local interactions, where the behavior of the system is significantly influenced by distant events or past observations [18]. For a detailed discussion, we refer to the monographs by Kilbas et al. [27], Lakshmikanthem et al. [28], Miller and Ross [29], Podlubny [31], Almeida et al. [11], and the references therein. Kilbas et al. [27] provided generalized Riemann-Liouville fractional operators to define the derivative with respect to a function. Several fractional derivatives concerning a function have been introduced, and their properties have been studied [5,8,9,14,24,34,37,38,39]. Katugampola expands upon the established framework of Erd'elyi-Kober operators, which are commonly used in the field of fractional derivatives. However, Katugampola introduced a specific modification and variation to these operators, which is referred to as the Hilfer-Katugampola fractional derivative [26]. These properties provide insights into the behavior of such fractional derivatives and their interactions with other mathematical operations.
Terminal value problems have applications in various fields where the ability to specify a desired final state is crucial to the problem at hand. For instance, in optimal control, terminal value problems are employed to determine the optimal control that minimizes a cost function over a given time frame. The terminal value serves as a constraint, allowing the specification of a particular final position or velocity for the system. Researchers have extensively investigated the existence and uniqueness of solutions to the terminal value problem, exploring the conditions under which a unique value can be determined for the system. For example, Benchohra et al. [16] investigated the existence and uniqueness of solutions for the terminal value problem in the Hilfer-Katugampola operator by using different types of classical fixed point theory. Almalahi et al. [6,7], developed sufficient conditions for the existence and uniqueness of solutions as well as stability results for the new terminal problems with respect to another function. Shah et al. [35] introduced some important notes on terminal value problems for fractional differential equations (FDEs) on infinite intervals. Abdo et al. [1] investigated the novel properties for the solution of a coupled system for terminal value problems involving a generalized Hilfer fractional operator. On the other hand, the symmetric solutions offer valuable insights into the behavior, stability, and simplification of mathematical and physical systems. They play a crucial role in simplifying analysis, facilitating physical interpretation, aiding in stability analysis, enhancing numerical computations, and contributing to the development of mathematical theories [42,44].
All of the aforementioned studies solely examined the properties of solutions within a single Banach space. However, our study seeks to expand upon this by investigating how the behavior of solutions is influenced by different weighted Banach spaces. Also, we aim to explore the impact of altering the function w on the solutions, while also examining multiple symmetric cases to observe the resulting variations in their behavior.
In this study, we investigate the existence and uniqueness of solutions for a system of terminal FDEs with w-Hilfer fractional derivatives and study their Ulam-Hyers (UH) stability [19,33,40]. Specifically, we study the fractional system
{Dpi,αa+,w(ι)ςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), ι∈(a,b],a>0,ςi(b)=υi∈R, i=1,2,....,n, | (1.1) |
where
● w∈C1([a,b],R) is an increasing function such that w′(ι)≠0, for all ι∈(a,b).
● Dpi,αa+,w(ι) denotes the w -Hilfer fractional derivatives of order pi,(0<pi<1) and type α,(0≤α≤1), for each i=1,2,....,n.
● The functions Fi:J×Rn→R,i=1,2,....,n, are continuous and satisfy some conditions that will be described later.
One of the primary motivations for studying terminal value problems with fractional derivatives is to analyze the stability properties of the corresponding dynamical systems. The w-Hilfer fractional derivative can capture complex behaviors such as subdiffusion or superdiffusion, which are not possible with classical integer-order derivatives. By formulating the problem in a weighted space, stability analysis can be performed in a manner that accounts for the importance or significance of different components or variables within the system. The innovative aspects of this work compared to the existing literature can be summarized as follows:
● We extend the analysis from single Banach spaces to different weighted Banach spaces, exploring how the change of weights affects solution properties. This allows us to capture additional structural information and understand the impact of weight functions on solutions.
● We investigate the sensitivity of solutions to changes in the weight function, providing insights into the relationship between the weighted functions and solution behavior by considering multiple symmetric cases to observe variations in solution behavior under different symmetry conditions.
● Furthermore, our study examines the system's properties in various Banach weighted spaces. We establish the existence, uniqueness, and stability of solutions for the system of w-Hilfer FDEs with minimal hypotheses, contributing to the theoretical understanding and providing a foundation for further analysis and applications.
● To the best of our knowledge, this is the first study to establish the existence, uniqueness, and stability of solutions for a system of w-Hilfer FDEs in different Banach-weighted spaces. Our approach expands the understanding of the system's behavior and broadens the range of potential applications.
● Additionally, the study addresses various parallel problems by considering specific function instances w. System (1.1) encompasses a fractional derivative that integrates several traditional fractional derivatives. This broadens the scope of the analysis and allows for a comprehensive investigation of the system's properties.
Notations: To enhance the readability, we fix the above notations and the following ones, we use these without any further mention.
(1) For 0<a<b, let J:=(a,b]⊂R.
(2)Γ(p) is the standard Gamma function given in [15].
(3)C(ˉJ,R) is the Banach space for all real-valued continuous functions defined on ˉJ, equipped with the norm ‖ς‖=maxι∈J|ς(ι)|, where ˉJ=[a,b]. The space ∏ni=1C(ˉJ,R) is the product space of n copies of C(ˉJ,R).
(4) For each i=1,2,...,n, let γi=pi+α(1−pi), such that pi∈(0,1) and α∈[0,1]. Then, we define the following spaces:
C1−γi,w(J,R)={ς:J→R;(w(ι)−w(a))1−γiςi(ι)∈C(ˉJ,R)}, |
and
Cγi1−γi,w(J,R)={ς∈C1−γi,w(J,R):Dγia+,wς∈C1−γi,w(J,R)}. |
The spaces C1−γi,w(J,R) are called weighted spaces. Clearly, these are Banach spaces equipped with the norm
‖ςi‖1−γi,w=maxι∈J|(w(ι)−w(a))1−γiςi(ι)|. |
(5)B=∏ni=1C1−γi,w(J,R) is the product space of C1−γi,w(J,R),i=1,2,....,n. The product spaces B are Banach spaces equipped with the norm
‖(ς1,ς2,.......,ςn)‖B=n∑i=1‖ςi‖1−γi. |
Hypotheses: To obtain our results, the following hypotheses must be satisfied:
(Hy1) For each (ι,ς1,ς2,....,ςn),(ι,ς∗1,ς∗2,....,ς∗n)∈J×Rn, there exist constants κjFi>0,i,j=1,2,....,n such that
|Fi(ι,ς1(ι),ς2(ι),....,ςn(ι))−Fi(ι,ς∗1(ι),ς∗2(ι),....,ς∗n(ι))|≤n∑j=1κjFi|ςj(ι)−ς∗j(ι)|. |
(Hy2) For each (ι,ς1,ς2,....ςn)∈J×Rn, there exist φjFi>0 with
|Fi(ι,ς1(ι),ς2(ι),....ςn(ι))|≤n∑j=1φjFi|ςj(ι)|. |
This paper is organized as follows. In Section 2, we state the needed fundamental notions of fractional derivatives, lemmas, and theorems that will be used in the conversion of the system into a fixed-point problem. In Section 3, we study the existence and uniqueness results of system (1.1), and discuss some properties of the symmetric cases of system (1.1). In Section 4, we investigate the UH stability of the solutions of system (1.1) and its symmetric cases. To illustrate our findings, Section 5 exhibits numerical examples with four cases and provides graphs of the solutions for all cases. We end the paper with concluding remarks.
In this section, the background definitions, lemmas, and theorems are given [8,27,30,37].
Definition 2.1. Let 0<p<1, ς be a continuous function on [a,b], and w be an increasing function on (a,b) such that w′(ι)≠0 for all ι∈(a,b). Then
(1) The w-Riemann-Liouville fractional integral for ς of order p is given as
Ipa+,w(ι)ς(ι)=1Γ(p)∫ιaw′(s)(w(ι)−w(s))p−1ς(s)ds. |
(2) The Riemann-Liouville fractional derivative for ς of order p with respect to w is given as
Dpa+,w(ι)ς(ι)=(1w′(ι)ddι)I1−pa+,w(ι)ς(ι). |
Definition 2.2. Let 0<p<1, 0≤α≤1, ι>a, and w be an increasing function on (a,b) such that w′(ι)≠0 for all ι∈(a,b). Then, the w−Hilfer fractional derivative of ς of order p and type α is defined by
Dp,αa+,w(ι)ς(ι)=Iα(1−p)a+,w(ι)(1w′(ι)ddι) I(1−α)(1−p)a+,w(ι) ς(ι). | (2.1) |
Clearly,
Dp,αa+,w(ι)ς(ι)=Iγ−pa+,w(ι) Dγa+,w(ι)ς(ι), |
where γ=p+α(1−p).
Lemma 2.3. For p,η,δ>0, we have
(1) Ipa+,w(ι) Iηa+,w(ι)ς(ι)=Ip+ηa+,w(ι)ς(ι).
(2) Ipa+,w(ι)(w(ι)−w(a))δ−1=Γ(δ)Γ(p+δ)(w(ι)−w(a))p+δ−1.
Note that Dγa+,w(ι)(w(ι)−w(a))γ−1=0, where γ=p+α(1−p).
Lemma 2.4. Let ς∈C(J,R), p∈(0,1) and α∈[0,1]. For γ=p+α(1−p), we have
Ipa+,w(ι) Dpa+,w(ι)ς(ι)=ς(ι)−(w(ι)−w(a))γ−1Γ(γ) I1−γa+,w(ι)ς(ι). |
Lemma 2.5. For each 0<p<1,0≤α≤1,γ=p+α(1−p), and ς∈Cγ1−γ;w(ι)(J,R), we have
Dp,αa+,w(ι)Ipa+,w(ι)ς(ι)=ς(ι), Iγa+,w(ι)Dγa+,w(ι)ς(ι)=Ipa+,w(ι) Dp,αa+,w(ι)ς(ι), |
and
Dγa+,w(ι)Ipa+,w(ι)ς(ι)=Dα(1−p)a+,w(ι)ς(ι). |
Definition 2.6. (Contraction mapping) Let (X,d) be a metric space, and Q:X→X be a mapping. Then, Q is a contraction mapping if there exists a constant 0≤L<1 such that for all ς,y∈X, the following inequality holds:
d(Q(ς),Q(y))≤L d(ς,y). |
The constant L is referred to as the contraction constant. If L=0, the mapping is called a strict contraction.
Theorem 2.7. (Banach fixed point theorem) Let (X,d) be a complete metric space, and Q:X→X be a contraction mapping. Then, the mapping Q has a unique fixed point in X.
Definition 2.8. Let a<b be real numbers, and X be the space of all continuous function h on [a,b]. The set Ω⊂X is said to be
(1) a uniformly bounded set if there exists M>0 with ‖h‖=supι∈[a,b]|h(ι)|<M, for every h∈Ω.
(2) an equicontinuous set if for every ε>0 there exists δ>0 such that
|ι−ι∗|<δ⟹|h(ι)−h(ι∗)|<ε, |
for all h∈Ω and for any ι,ι∗∈[a,b].
Theorem 2.9. (Arzelá-Ascoli's theorem) Let a<b be real numbers, and X be the space of all continuous function on [a,b]. The set Ω⊂X is relatively compact if it is a uniformly bounded and equicontinuous set.
Theorem 2.10. (Schauder's fixed-point theorem) Let X be a Banach space and Ω⊂X be a non-empty, compact, convex subset. If Q:Ω→Ω is a continuous mapping, and Q(Ω) is a relatively compact subset of X, then Q has at least one fixed point in Ω.
In this section, we establish and develop sufficient conditions for the existence and uniqueness as well as UH stability results for the system (1.1). We begin by stating an equivalent integrals for the considered system.
In the following theorem, we convert system (1.1) into equivalent integral equations.
Theorem 3.1. Let 0<pi<1, 0≤α≤1, and γi=pi+α(1−pi),i=1,2,....,n. If Fi(⋅,ς1(⋅),ς2(⋅),......,ςn(⋅))∈C1−γ,w(J,R), and ςi∈Cγi1−γi,w(J,R), then ςi satisfies system (1.1) if and only if ςi is given by:
ςi(ι)=(w(ι)−w(a)w(b)−w(a))γi−1[υi−Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), | (3.1) |
where
Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)=1Γ(pi)∫baw′(s)(w(b)−w(s))pi−1Fi(s,ς1(s),..,ςn(s))ds, |
and
Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))=1Γ(pi)∫ιaw′(s)(w(ι)−w(s))pi−1Fi(s,ς1(s),....,ςn(s))ds. |
Proof. Let ςi∈Cγi1−γi,w(J,R) be a solution of system (1.1). We prove that ςi is also a solution of Eq (3.1). From the definition of Cγi1−γi,w(J,R), and using the definition 2.2, we have
I1−γia+,w(ι)ςi(ι)∈C1−γi,w(J,R) and Dγia+,w(ι)ςi(ι)=D1a+,w(ι)I1−γia+,w(ι)ςi(ι)∈C1−γi,w(J,R). |
By the definition of the space Cn1−γi,w(J,R), it follows that
I1−γia+,w(ι)ςi(ι)∈C11−γi,w(J,R). | (3.2) |
Using Lemma 2.4, with p=γi, we obtain
(Iγia+,w(ι) Dγia+,w(ι)ςi)(ι)=ςi(ι)−(w(ι)−w(a))γi−1Γ(γi) limι→a(I1−γia+,w(ι)ςi)(ι). | (3.3) |
By hypothesis (ςi∈Cγi1−γi,w(J,R)) and Lemma 2.5, we have
(Iγia+,w(ι) Dγia+,w(ι)ςi)(ι)=Ipia+,w(ι) Dpi,αa+,w(ι)ςi(ι). |
Since Dpi,αa+,w(ι)ςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), we have
(Iγia+,w(ι) Dγia+,w(ι)ςi)(ι)=Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). | (3.4) |
Comparing Eqs (3.3) and (3.4), we see that
ςi(ι)=(w(ι)−w(a))γi−1Γ(γi)limι→a(I1−γia+,w(ι)ςi)(ι)+Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). | (3.5) |
Replacing ι with b in Eq (3.5), we get
ςi(b)=(w(b)−w(a))γi−1Γ(γi)limι→a(I1−γia+,w(ι)ςi)(ι)+Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). |
Using the terminal condition ςi(b)=υi, we get
(w(b)−w(a))γi−1Γ(γi)limι→a(I1−γia+,w(ι)ςi)(ι)+Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))=υi. |
Thus,
limι→a(I1−γia+,w(ι)ςi)(ι)=Γ(γi)(w(b)−w(a))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))]. | (3.6) |
Putting Eq (3.6) into Eq (3.5), we get
ςi(ι)=((w(ι)−w(a))(w(b)−w(a)))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))]+Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). |
Hence, ςi(ι) satisfies Eq (3.1).
Conversely, let ςi∈Cγi1−γi,w(J,R) be functions satisfying (3.1). We prove that ςi is also a solution of the system (1.1). Applying the operator Dγia+,w(ι) on both sides of Eq (3.1), we get
Dγia+,w(ι)ςi(ι)=Dγia+,w(ι)((w(ι)−w(a))(w(b)−w(a)))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))]+Dγia+,w(ι)Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). | (3.7) |
By Lemma 2.3, we have Dγia+,w(ι)(w(ι)−w(a))γi−1=0, and hence Eq (3.7) becomes
Dγia+,w(ι)ςi(ι)=Dγia+,w(ι)Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). | (3.8) |
Then, from Lemma 2.5, the right side of Eq (3.8) becomes
Dγia+,w(ι)Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))=Dα(1−pi)a+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). |
Thus, Eq (3.8) becomes
Dγia+,w(ι)ςi(ι)=Dα(1−pi)a+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). |
From Eq (3.2), we have Dγia+,w(ι)ςi(ι)∈C1−γi,w(J,R), and hence, Eq (3.8) implies
Dγia+,w(ι)ςi(ι)=Dα(1−pi)a+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))∈C1−γi,w(J,R). | (3.9) |
As Fi(⋅,ς1(⋅),ς2(⋅),......,ςn(⋅))∈C1−γi,w(J,R), it follows that
I1−α(1−pi)a+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))∈C1−γi,w(J,R). | (3.10) |
From Eqs (3.9) and (3.10) and the definition of the space Cn1−γi,w(J,R), we get
I1−α(1−pi)a+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))∈C11−γi,w(J,R). |
Now, by applying the operator Iα(1−pi)a+,w(ι) on both sides of Eq (3.9) and using Lemma 2.4, we have
Iα(1−pi)a+,w(ι)Dγia+,w(ι)ςi(ι)=ςi(ι)−I1−α(1−pi)a+,w(ι)Fi(a,ς1(a),ς2(a),......,ςn(a))Γ(α(1−pi))(w(ι)−w(a))α(1−pi)−1=Dpi,αa+,w(ι)ςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). | (3.11) |
From Eq (2.1), Eq (3.11) reduces to
Dpi,αa+,w(ι)ςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). |
Thus, Eq (1.1) holds.
Let Q:B→B be an operator defined by Q=(Q1,Q2,.....,Qn), where
Qi(ς1,ς2,....,ςn)(ι)=(w(ι)−w(a)w(b)−w(a))γi−1[υi−Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)). |
Theorem 3.2. Assume both (Hy1) and (Hy2) hold. For r=1,2,.....,n, let
Pr=n∑i=12(w(b)−w(a))1−γiφrFiΓ(γr)Γ(pi+γr)(w(b)−w(a))pi+γr−1, |
and
ζr=n∑i=12(w(b)−w(a))1−γiκrFiΓ(γr)Γ(pi+γr)(w(b)−w(a))pi+γr−1. |
If ζ=maxr{ζr}<1 and P=maxr{Pr}<1, then system (1.1) has at least one solution.
Proof. Let β≥11−P∑ni=1|υi|(w(b)−w(a))γi−1, and define a closed ball Sβ as
Sβ={(ς1,ς2,....ςn)∈B:‖(ς1,ς2,....ςn)‖B≤β}. |
To apply Schauder's fixed point theorem, we divide the proof into the following steps.
Step 1: Q(Sβ)⊂Sβ. For any (ς1,ς2,....ςn)∈Sβ, we have
‖Q(ς1,ς2,....ςn)‖B≤n∑i=1‖Qi(ς1,ς2,....ςn)‖C1−γi,w. | (3.12) |
From (Hy2), we get
|Qi(ς1,ς2,....ςn)(ι)|≤(w(ι)−w(a)w(b)−w(a))γi−1[|υi|+Ipia+,w(b)|Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|]+Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))≤(w(ι)−w(a)w(b)−w(a))γi−1[|υi|+ Ipia+,w(b)(n∑j=1φjFi|ςj(b)|)]+Ipia+,w(ι)(n∑j=1φjFi|ςj(ι)|). | (3.13) |
Taking into consideration that (ς1,ς2,....ςn)∈B, we have
Ipia+,w(ι)(n∑j=1φFji|ςj(ι)|)=n∑j=1Ipia+,w(ι)φjFi(w(ι)−w(a))γj−1|(w(ι)−w(a))1−γj(ςj(ι))|≤n∑j=1φjF1Γ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςj‖C1−γj,w. | (3.14) |
From (3.13) and (3.14), it follows that
|(w(ι)−w(a))1−γiQi(ς1,ς2,....ςn)(ι)|≤(w(b)−w(a))1−γi[|υi|+n∑j=1φjFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςj‖C1−γj,w]+(w(ι)−w(a))1−γin∑j=1φjFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςj‖C1−γj,w. |
Since γi<1, then 1−γi>0, and hence (w(ι)−w(a))1−γi<(w(b)−w(a))1−γi. Then, we get
‖Qi(ς1,ς2,....ςn)‖C1−γi,w≤|υi|(w(b)−w(a))1−γi+2(w(b)−w(a))1−γi×n∑j=1φjFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςj‖C1−γj,w. |
Thus,
‖Q(ς1,ς2,....ςn)‖B≤n∑i=1‖Qi((ς1,ς2,....ςn)‖C1−γi,w≤n∑i=1|υi|(w(b)−w(a))1−γi+n∑i=12(w(b)−w(a))1−γi×n∑j=1φjFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςj‖C1−γj,w≤n∑i=1|υi|(w(b)−w(a))1−γi+P‖(ς1,ς2,....ςn)‖B≤β. |
Hence, Q(Sβ)⊂Sβ.
Step 2: Q is continuous. Let (ς1k,ς2k,....,ςnk) be a sequence in Sβ such that (ς1k,ς2k,....,ςnk)→(ς1,ς2,....,ςn) in Sβ as k→∞. Then, we have
‖Q(ς1k,ς2k,....,ςnk)(ι)−Q(ς1,ς2,....,ςn)(ι)‖B≤n∑i=1‖(Qi(ς1k,ς2k,....,ςnk)−Qi(ς1,ς2,....,ςn))(ι)‖C1−γi,w≤n∑i=1(w(b)−w(a))1−γiIpia+,w(ι)|Fi(ι,ς1k(ι),ς2k(ι),....,ςnk(ι))(b)−Fi(ι,ς1(ι),ς2(ι),....,ςn(ι))(b)|+n∑i=1(w(ι)−w(a))1−γiIpia+,w(ι)|Fi(ι,ς1k(ι),ς2k(ι),....,ςnk(ι))−Fi(ι,ς1(ι),ς2(ι),....,ςn(ι))|≤n∑i=12(w(b)−w(a))1−γin∑j=1κiFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςjk−ςj‖C1−γj,w≤ζ‖(ς1k,ς2k,....,ςnk)−(ς1,ς2,....ςn)‖B. |
This implies that ‖Q(ς1k,ς2k,....,ςnk)−Q(ς1,ς2,....,ςn)‖B→0 as k→∞. So, Q is continuous.
Step 3: Q(Sβ)is equicontinuous.
For ι1,ι2∈J with ι1<ι2, and for any (ς1,ς2,....,ςn)∈Sβ, we have
|Q(ς1,ς2,....,ςn)(ι2)−Q(ς1,ς2,....,ςn)(ι1)|≤n∑i=1|Qi(ς1,ς2,....,ςn)(ι2)−Qi(ς1,ς2,....,ςn)(ι1)|≤n∑i=1(w(ι2)−w(a))γi−1−(w(ι1)−w(a))γi−1(w(b)−w(a))γi−1×[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+n∑i=1|Ipia+,w(ι2)Fi(ι2,ς1,ς2,....,ςn)−Ip1a+,w(ι1)Fi(ι1,ς1,ς2,....,ςn)|. |
Since Fi(⋅,ς1(⋅),....,ςn(⋅)) are continuous on J, there exist ξFi∈R such that
|Fi(⋅,ς1(⋅),....,ς1(⋅))|≤ξFi. |
Hence,
|Q(ς1,ς2,....,ςn)(ι2)−Q(ς1,ς2,....,ςn)(ι1)|≤n∑i=1(w(ι2)−w(a))γi−1−(w(ι1)−w(a))γi−1(w(b)−w(a))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+n∑i=1|Ipia+,w(ι2)Fi(ι2,ς1(ι2),ς2(ι2),....,ςn(ι2))−Ipia+,w(ι1)Fi(ι1,ς1(ι1),ς2(ι1),....,ςn(ι1))|≤n∑i=1(w(ι2)−w(a))γi−1−(w(ι1)−w(a))γi−1(w(b)−w(a))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+n∑i=11Γ(pi)∫ι1aw′(ι)[(w(ι2)−w(ι))pi−1−(w(ι1)−w(ι))pi−1]|Fi(ι,ς1(ι),ς2(ι),....,ςn(ι))|dι+n∑i=11Γ(pi)∫ι2ι1w′(ι)(w(ι2)−w(ι))pi−1|Fi(ι,ς1(ι),ς2(ι),....,ςn(ι))|dι≤n∑i=1(w(ι2)−w(a))γi−1−(w(ι1)−w(a))γi−1(w(b)−w(a))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+n∑i=1ξFiΓ(pi)∫ι1aw′(ι)[(w(ι2)−w(ι))pi−1−(w(ι1)−w(ι))pi−1]dι+n∑i=1ξFiΓ(pi)∫ι2ι1w′(ι)(w(ι2)−w(ι))pi−1dι≤n∑i=1(w(ι2)−w(a))γi−1−(w(ι1)−w(a))γi−1(w(b)−w(a))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+n∑i=1ξFiΓ(pi+1)[(w(ι2)−w(a))pi−(w(ι1)−w(a))pi]. | (3.15) |
Thus, we get
|Q(ς1,ς2,....,ςn)(ι2)−Q(ς1,ς2,....,ςn)(ι1)|≤n∑i=1(w(ι2)−w(a))γi−1−(w(ι1)−w(a))γi−1(w(b)−w(a))γi−1[υi−Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))(b)]+n∑i=12ξFiΓ(pi+1)[(w(ι2)−w(a))pi−(w(ι1)−w(a))pi]. |
Since w(ι) is an increasing function, then
|Q(ς1,ς2,....,ςn)(ι2)−Q(ς1,ς2,....,ςn)(ι1)|→0, as ι2→ι1. |
This implies that Q(Sβ) is equicontinuous. Since Q(Sβ) is also uniformly bounded, then by Arzelá-Ascoli's theorem, it is relatively compact. By Theorem 3.1, Q has a fixed point. Consequently, system (1.1) possesses at least one solution.
Theorem 3.3. Suppose that hypothesis (Hy1) holds. For r=1,2,...,n, let
ζr=n∑i=12(w(b)−w(a))1−γiκrFiΓ(γr)Γ(pi+γr)(w(b)−w(a))pi+γr−1. |
If ζ=maxr{ζr}<1, then system (1.1) has a unique solution.
Proof. To show that system (1.1) has a unique solution, we show that the operator Q is a contraction, where Φ as defined in Theorem 3.1.
For all ι∈J and (ς1,ς2,....,ςn),(ς∗1,ς∗2,....,ς∗n)∈B, we have
‖Q(ς1,ς2,....,ςn)−Q(ς∗1,ς∗2,....,ς∗n)‖B≤n∑i=1‖Qi(ς1,ς2,....,ςn)−Qi(ς∗1,ς∗2,....,ς∗n)‖C1−γi,w≤n∑i=1(w(b)−w(a))1−γiIpia+,w(b)|Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))−Fi(ι,ς∗1(ι),ς∗2(ι),....,ς∗n(ι))|+n∑i=1(w(ι)−w(a))1−γiIpia+,w(ι)|Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))−Fi(ι,ς∗1(ι),ς∗2(ι),....,ς∗n(ι))|≤n∑i=12(w(b)−w(a))1−γi(n∑j=1κjFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−1‖ςj−ς∗j‖C1−γj,w)≤ζ‖(ς1,ς2,....,ςn)−(ς∗1,ς∗2,....,ς∗n)‖B. |
Since ζ<1, the operator Q is a contraction map. Thus, by Theorem 2.7, system (1.1) has a unique solution.
In this subsection, we consider some symmetric cases of system (1.1):
Case1: w(ι)=ι. In this case, system (1.1) reduces to a system of Hilfer fractional equations in the following form
{Dpi,αa+,ιςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), ι∈J,ςi(b)=υi∈R,i=1,2,....,n, | (3.16) |
where Dpi,αa+,ι represents the Hilfer fractional derivative of order pi [23]. Let B1=∏ni=1C1−γi,ι(J,R), where C1−γi,ι is the weighted space defined as in the notation part and equipped with the norm
‖ςi‖1−γi=maxι∈J|(ι−a)1−γiςi(ι)|. | (3.17) |
Theorems 3.2 and 3.3 imply the following corollary.
Corollary 3.4. Assume that both (Hy1) and (Hy2) hold. For 1≤r≤n, let
P⋆r=n∑i=12(b−a)γi−1φrFiΓ(γr)Γ(pi+γr)(b−a)pi+γr−1, | (3.18) |
and
ζ⋆r=n∑i=12(b−a)γi−1κrFiΓ(γr)Γ(pi+γr)(b−a)pi+γr−1. | (3.19) |
If P⋆=maxr{P⋆r}<1, and ζ⋆=maxr{ζ⋆r}<1, then system (3.16) has at least one solution (ς1,ς2,....,ςn)∈B1, where ςi is given by
ςi(ι)=(ι−a)γi−1(b−a)γi−1[υi−Ipia+,bFi(ι,ς1(ι),ς2(ι),......,ςn(ι))]+Ipia+,ιFi(ι,ς1(ι),ς2(ι),......,ςn(ι)), |
where
Ipia+,bFi(ι,ς1(ι),ς2(ι),....,ςn(ι))=1Γ(pi)∫ba(b−s)pi−1Fi(s,ς1(s),ς2(s),....,ςn(s))ds, |
and
Ipia+,ιFi(ι,ς1(ι),ς2(ι),....,ςn(ι))=1Γ(pi)∫ιa(ι−s)pi−1Fi(s,ς1(s),ς2(s),....,ςn(s))ds. |
Case 2: w(ι)=logι. In this case, system (1.1) has the form
{Dpi,α1+,logιςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), ι∈J,a=1,ςi(b)=υi∈R,i=1,2,....,n, | (3.20) |
where, Dpi,α1+,logι is the Hilfer-Hadamard fractional derivative of order pi [22,32]. Let B2=∏ni=1C1−γi,logι(J,R), where C1−γi,logι is equipped with the norm
‖ςi‖1−γi,logι=maxι∈J|(logι)1−γiςi(ι)|. | (3.21) |
Theorems 3.2 and 3.3 then imply the following corollary.
Corollary 3.5. Assume that both (Hy1) and (Hy2) hold. For 1≤r≤n, let
P∗∗r=n∑i=12(logι)γi−1φrFiΓ(γr)Γ(pi+γr)(logι)pi+γr−1, | (3.22) |
and
ζ∗∗j=2(logι)γi−1n∑j=1κjFiΓ(γj)Γ(pi+γj)(logι)pi+γj−1. | (3.23) |
If P∗∗=maxr{P∗∗r}<1, and ζ∗∗=maxr{ζ∗∗r}<1, then, system (3.20) has at least one solution (ς1,ς2,....,ςn)∈B2, where ςi is given by
ςi(ι)=(logι)γi−1(logb)γi−1[υi−Ipi1+,logbFi(ι,ς1(ι),ς2(ι),......,ςn(ι))]+Ipi1+,logιFi(ι,ς1(ι),ς2(ι),......,ςn(ι)), |
where
Ipi1+,logbFi(ι,ς1(ι),ς2(ι),....,ςn(ι))=1Γ(pi)∫b11s(logbs)pi−1Fi(s,ς1(s),ς2(s),....,ςn(s))ds, |
and
Ipi1+,logιFi(ι,ς1(ι),ς2(ι),....,ςn(ι))=1Γ(pi)∫ι11s(logιs)pi−1Fi(s,ς1(s),ς2(s),....,ςn(s))ds. |
Case 3: w(ι)=ιq, where q>0. In this case, system (1.1) becomes a system of Hilfer-Katugampola in the following form.
{Dpi,αa+,ιqςi(ι)=Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), ι∈J,a>0,ςi(b)=υi∈R,i=1,2,....,n, | (3.24) |
where Dpi,αa+,ιp is the Hilfer-Katugampola fractional derivative of order pi [26,30]. Let B3=∏ni=1C1−γi,ιq(J,R), where C1−γi,ιq is the weighted space equipped with the norm
‖ςi‖1−γi,ιq=maxι∈J|(ιq−aq)1−γiςi(ι)|. | (3.25) |
Theorems 3.2 and 3.3 imply the following corollary.
Corollary 3.6. Assume that (Hy1) and (Hy2) hold. For 1≤r≤n, let
P∗∗∗r=n∑i=12(bq−aq)γi−1φrFiΓ(γr)Γ(pi+γr)(bq−aq)pi+γr−1, | (3.26) |
and
ζ∗∗∗r=n∑i=12(bq−aq)γi−1κrFiΓ(γr)Γ(pi+γr)(bq−aq)pi+γr−1. | (3.27) |
If P∗∗∗=maxr{P∗∗∗r}<1, and ζ∗∗∗=maxr{ζ∗∗∗r}<1, then system (3.20) has at least one solution (ς1,ς2,....,ςn)∈B3, where ςi is given by
ςi(ι)=(ιq−aqq)γi−1(bq−aqq)γi−1[υi−Ipia+,bqFi(ι,ς1(ι),ς2(ι),......,ςn(ι))]+Ipia+,ιqFi(ι,ς1(ι),ς2(ι),......,ςn(ι)), |
where
Ipia+,bqFi(ι,ς1(ι),ς2(ι),....,ςn(ι))=1Γ(pi)∫baqsq−1(bq−sq)pi−1Fi(s,ς1(s),ς2(s),....,ςn(s))ds, |
and
Ipia+,ιqFi(ι,ς1(ι),ς2(ι),....,ςn(ι))=1Γ(pi)∫ιaqsq−1(ιq−sq)pi−1Fi(s,ς1(s),ς2(s),....,ςn(s))ds. |
Remark 3.7. There are many symmetric cases of the function w and parameter α that result in symmetric systems, some of which have already been discussed in the literature. For example.
1) For w(ι)=ι and α=0, system (1.1) is equivalent to the Riemann-Liouville system [27].
2) For w(ι)=ι and α=1, system (1.1) is equivalent to the Caputo system [27].
3) For w(ι)=ιq and α=1, system (1.1) is equivalent to the Caputo-Katugampola system [10].
4) For w(ι)=ιq and α=0, system (1.1) is equivalent to the Katugampola system [26].
5) For w(ι)=ι and α∈(0,1), system (1.1) is equivalent to the Hilfer system [23].
6) For w(ι)=ιq,q>0 and α∈(0,1) system (1.1) is equivalent to the Hilfer-Katugampola system [30].
7) For w(ι)=logι and α∈(0,1), system (1.1) is equivalent to the Hilfer-Hadamard system [32].
8) For w(ι)=logι and α=0), system (1.1) is equivalent to the Hadamard system [22].
9) For w(ι)=logι and α=1, system (1.1) is equivalent to the Caputo-Hadamard system [25].
The UH stability of system (1.1) is discussed in this section. Following the notation of this paper, we state the following definition for UH for system (1.1) [19,33,40].
Definition 4.1. System (1.1) is UH stable if there exists a positive constant Υ=max{Υ1,Υ2,...,Υn} such that for some ε=max{ε1,ε2,...,εn}>0 and for each solution (z1,z2,....,zn) of system (1.1) satisfying the following inequality
|Dpi,αa+,w(ι)zi(ι)−Fi(ι,z1(ι),z2(ι),......,zn(ι))|≤εi, | (4.1) |
there exists a unique solution (ς1,ς2,....,ςn) for system (1.1) such that
‖(z1,z2,....,zn)−(ς1,ς2,....,ςn)‖B≤Υε. | (4.2) |
Remark 4.2. The functions (z1,z2,....,zn) are a solution of inequality (4.1) if and only if there exist functions σi(ι)∈C1−γi,w,i=1,2,....,n (which depends on solution zi) such that
● |σi(ι)|≤εi.
● Dpi,αa+,w(ι)zi(ι)=Fi(ι,z1(ι),z2(ι),......,zn(ι))+σi(ι).
Theorem 4.3. Under hypothesis (Hy1), system (1.1) is UH stable, provided that
2κiFiΓ(γi)Γ(pi+γi)(w(b)−w(a))pi<1, |
for all 1≤i≤n.
Proof. Let ε=max{ε1,ε2,....,εn}>0 and let (z1,z2,....,zn)∈B be a solution of system (1.1) satisfying inequality (4.1), and let (ς1,ς2,....,ςn)∈B be the unique solution of system (1.1). By virtue of Theorem 3.1, we obtain
ςi(ι)=Aςi−((w(ι)−w(a))(w(b)−w(a)))γi−1Ipia+,w(b)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))+Ipia+,w(ι)Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)), | (4.3) |
where
Aςi=υi((w(ι)−w(a))(w(b)−w(a)))γi−1, | (4.4) |
since, we have assumed that (z1,z2,....,zn)∈B is a solution of system (1.1) satisfying the inequality (4.1). Then, by Remark 4.2, we conclude that (z1,z2,....,zn) satisfies the following system
{Dpi,αa+,w(ι)zi(ι)=Fi(ι,z1(ι),z2(ι),......,zn(ι))+σi(ι), ι∈J,zi(b)=υi∈R,i=1,2,....,n. | (4.5) |
By Theorem 3.1, the solution zi(ι) of system (4.5) is given as
zi(ι)=Azi−((w(ι)−w(a))(w(b)−w(a)))γi−1Ipia+,w(b)[Fi(ι,z1(ι),z2(ι),...,zn(ι))+σi(ι)]+Ipia+,w(ι)[Fi(ι,z1(ι),z2(ι),...,zn(ι))+σi(ι)], |
where
Azi=υi((w(ι)−w(a))(w(b)−w(a)))γi−1. | (4.6) |
Hence, we have
|zi(ι)−ςi(ι)|≤|Azi−Aςi|+((w(ι)−w(a))(w(b)−w(a)))γi−1Ipia+,w(b)|σi(ι)|+Ipia+,w(ι)|σi(ι)|+((w(ι)−w(a))(w(b)−w(a)))γi−1Ipia+,w(b)|Fi(ι,z1(ι),z2(ι),......,zn(ι))−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|+Ipia+,w(ι)|Fi(ι,z1(ι),z2(ι),......,zn(ι))−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|. | (4.7) |
By (4.4) and (4.6), we have Azi=Aςi. Since the function w is increasing for each ι∈J and γi<1 for i=1,2,...n, we get
((w(ι)−w(a))(w(b)−w(a)))γi−1<1. |
Hence, Eq (4.7) becomes
|zi(ι)−ςi(ι)|≤Ipia+,w(b)|σi(ι)|+Ipia+,w(ι)|σi(ι)|+Ipia+,w(b)|Fi(ι,z1(ι),z2(ι),......,zn(ι))−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|+Ipia+,w(ι)|Fi(ι,z1(ι),z2(ι),......,zn(ι))−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|. |
By the first part of Remark 4.2 and (Hy1), we get
|zi(ι)−ςi(ι)|≤εi2(w(b)−w(a))piΓ(pi+1)+2Ipia+,w(b)n∑j=1κjFi|zj(ι)−ςj(ι)|. |
Thus, for i=1, we have
|z1(ι)−ς1(ι)|≤ε12(w(b)−w(a))p1Γ(p1+1)+2Ip1a+,w(b)n∑j=1κjF1|zj(ι)−ςj(ι)|. |
Since ‖ςi‖1−γi,w=maxι∈J|(w(ι)−w(a))1−γiςi(ι)|, and w is increasing function, then
‖z1−ς1‖1−γ1,w≤ε12(w(b)−w(a))p1Γ(p1+1)+2n∑j=1(w(b)−w(a))1−γ1Ip1a+,w(ι)κjF1|zj(ι)−ςj(ι)|≤ε12(w(b)−w(a))p1Γ(p1+1)+2κ1F1Γ(γ1)Γ(p1+γ1)(w(b)−w(a))p1‖z1−ς1‖C1−γ1,w+2n∑j=2κjF1Γ(γj)Γ(p1+γj)(w(b)−w(a))p1+γj−γ1‖zj−ςj‖C1−γj,w. |
Hence, we have
(1−2κ1F1Γ(γ1)Γ(p1+γ1)(w(b)−w(a))p1)‖z1−ς1‖1−γ1,w≤ε12(w(b)−w(a))p1+1−γ1Γ(p1+1)+2n∑j=2κjF1Γ(γj)Γ(p1+γj)(w(b)−w(a))p1+γj−γ1‖zj−ςj‖C1−γj,w≤ε12(w(b)−w(a))p1+1−γ1Γ(p1+1)+2ε1n∑j=2κjF1Γ(γj)Γ(p1+γj)(w(b)−w(a))p1+γj−γ1‖zj−ςj‖C1−γj,w. |
Note that the positive values of the Gamma function on (0,1] are necessary to obtain the last inequality.
Let c1=1−2κ1F1Γ(γ1)Γ(p1+γ1)(w(b)−w(a))p1, and
Υ1=2(w(b)−w(a))p1Γ(p1+1)+2n∑j=2κjF1Γ(γj)Γ(p1+γj)(w(b)−w(a))p1+γj−γ1‖zj−ςj‖C1−γj,w. |
If 2κ1F1Γ(γ1)Γ(p1+γ1)(w(b)−w(a))p1<1, then c1>0. Then, we obtain
‖z1−ς1‖1−γ1,w≤ε1c1Υ1. |
Similarly, for each i=2,3,..,n, provided that 2κiFiΓ(γi)Γ(pi+γi)(w(b)−w(a))pi<1, we have
‖zi−ςi‖1−γi,w≤εiciΥi, |
where ci=1−2κiFiΓ(γi)Γ(pi+γi)(w(b)−w(a))pi, and
Υi=2(w(b)−w(a))piΓ(pi+1)+n∑j≠iκjFiΓ(γj)Γ(pi+γj)(w(b)−w(a))pi+γj−γi‖zj−ςj‖C1−γj,w. |
Therefore, if 2κiFiΓ(γi)Γ(pi+γi)(w(b)−w(a))pi<1,i=1,2,..n, then
‖(z1,z2,....,zn)−(ς1,ς2,....,ςn)‖B≤n∑i=1‖zi−ςi‖1−γi,w≤n∑i=1εiciΥi≤εΥ, |
where ε=maxi{εi} and Υ=∑ni=1Υici. By Definition 4.1, the solution of system (1.1) is UH stable.
According to Theorem 4.3, we can easily prove the UH stability of symmetric systems as follows.
(1) Under hypothesis (Hy1), system (3.16) is UH stable, provided that
2κiFiΓ(γi)Γ(pi+γi)(b−a)pi<1, |
for all 1≤i≤n.
(2) Under hypothesis (Hy1), system (3.20) is UH stable, provided that
2κiFiΓ(γi)Γ(pi+γi)(logba)pi<1, |
for all 1≤i≤n.
(3) Under the hypothesis (Hy1), the system (3.24) is UH stable, provided that
2κiFiΓ(γi)Γ(pi+γi)(bq−aq)pi<1, |
for all 1≤i≤n.
In this section, we consider an example of a system of w-Hilfer FDEs with different functions w.
Example 5.1. We study the system
{D12,120,w(ι)ςi(ι)=118(52ι)[1+∑nj=1|ςj(ι)|], ι∈(0,1],ςi(1)=3i2, i=1,2,...,n, | (5.1) |
for the cases w(ι)=eι,w(ι)=ι,w(ι)=logι, and w(ι)=ιq.
Here, pi=12,αi=12,γi=34,a=0,b=1,υi=3i2, and
Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))=118(52ι)[1+∑nj=1|ςj(ι)|]. |
Clearly, Fi(ι,ς1(ι),ς2(ι),......,ςn(ι)) are continuous functions on [0,1]×Rn and satisfy
|Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|≤118, |
and
|Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))−Fi(ι,z1(ι),z1(ι),......,zn(ι))|≤118n∑j=1|ςj−zj|, |
for all (ς1,ς2,....ςn),(z1,z2,....zn), i.e., φjFi=κjFi=118.
We fix the number of equations to be n=5, and investigate the following cases:
Case 1: w(ι)=eι. In this case, we have P=max{P1,...,P5}≃0.75<1, and ζ=max{ζ1,...,ζ5}=0.75<1, where Pr and ζr are given in Theorems 3.2 and 3.3, respectively.
Hence, all conditions in Theorem 3.2 and Theorem 3.3 are satisfied. Thus, the system has a unique solution (ς1,ς2,...,ς5). This solution has the form
ςi(ι)=(eι−1)−14(e−1)−14[3i2−1Γ(12)∫10es(e−es)−1218(52s)[1+∑5j=1|ςj(s)|]ds]+1Γ(12)∫ι0es(eι−es)−1218(52s)[1+∑5j=1|ςj(s)|]ds. |
On the other hand, let ι∈(0,1] and εi>0,i=1,2,...,n satisfying the following inequality
|D12,120+,eιςi(ι)−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|≤εi. |
Then, system (1.1) is UH stable with
‖(z1,z2,....,zn)−(ς1,ς2,....,ςn)‖B≤n∑i=1εiciΥi, |
where
Υi=2(eι−1)12Γ(12+1)+n∑j≠iκjFiΓ(γj)Γ(12+γj)(e−1)12+γj−γi>0, |
and
ci=1−2118Γ(34)Γ(12+34)(e−1)12=0.8>0. |
Symmetric cases:
Case 2: w(ι)=ι. In this case, we have P∗=max{P∗1,...,P∗5}≃0.37<1, and ζ∗=maxι∈J{ζ∗1,...,ζ∗5}=0.37<1, where P∗r and ζ∗r are given in (3.18) and (3.19), respectively.
Hence, all conditions in Theorem 3.2 and Theorem 3.3 are satisfied. Thus, the system has at least one solution (ς1,ς2,...,ς5) given by
ςi(ι)=ι−14[3i2−1Γ(12)∫10(1−s)−1218(52s)[1+∑5j=1|ςj(s)|]ds]+1Γ(12)∫ι0(ι−s)−1218(52s)[1+∑5j=1|ςj(s)|]ds. | (5.2) |
On the other hand, let ι∈(0,1] and εi>0,i=1,2,...,n, satisfying the following inequality,
|D12,120+,ιςi(ι)−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|≤εi. |
Then, system (3.16) is UH stable with
‖(z1,z2,....,zn)−(ς1,ς2,....,ςn)‖B≤n∑i=1εiciΥi, |
where
Υi=2Γ(12+1)+n∑j≠iκjFiΓ(γj)Γ(12+γj)>0, |
and
ci=1−2118Γ(34)Γ(12+34)=0.9>0. |
Case 3: w(ι)=logι. For this case, we need to change the interval J, as the function w is not defined on (0,1]. So, in this case, only a and b will be changed to be a=1,b=2, P∗∗=max{P∗∗1,...,P∗∗5}≃0.75<1, and ζ∗∗=max{ζ∗∗1,...,ζ∗∗5}=0.76<1, where P∗∗r and ζ∗∗r are defined in (3.22) and (3.23), respectively. Hence, all conditions in Theorem 3.2 and Theorem 3.3 are satisfied. Thus, the system has at least one solution (ς1,ς2,...,ς5) given by
ςi(ι)=(logι)−14(log2)−14[3i2−1Γ(12)∫21s−1(log2−logs)−1218(52s)[1+∑5j=1|ςj(s)|]ds]+1Γ(12)∫ι1s−1(logι−logs)−1218(52s)[1+∑5j=1|ςj(s)|]ds. | (5.3) |
On the other hand, let ι∈(1,2] and εi>0,i=1,2,...,n, satisfying the following inequality.
|D12,120+,logιςi(ι)−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|≤εi. |
Then, system (3.20) is UH stable with
‖(z1,z2,....,zn)−(ς1,ς2,....,ςn)‖B≤n∑i=1εiciΥi, |
where
Υi=2(log2−log1)12Γ(12+1)+n∑j≠iκjFiΓ(γj)Γ(12+γj)(log2−log1)12+γj−γi>0, |
and
ci=1−2118Γ(34)Γ(12+34)(log2−log1)12=0.9>0. |
Case 4: w(ι)=ιq,q>0. We also consider this case on (1,2], even though we can consider the system on [0,1], where it will be generalizing case 2. In this case, we have P=max{P∗∗∗1,..,P∗∗∗5}≃0.75<1, and ζ=max{ζ∗∗∗1,..,ζ∗∗∗5}=0.75<1, where P∗∗∗r and ζ∗∗∗r are defined in (3.26) and (3.27), respectively. Hence, all conditions in Theorem 3.2 are satisfied. Thus, the system has at least one solution (ς1,ς2,...,ςn) given by
ςi(ι)=(ιq−1)−14(2q−1)−14[3i2−1Γ(12)∫21qsq−1(2q−sq)−1218(52s)[1+∑5j=1|ςj(s)|]ds]+1Γ(12)∫ι1qsq−1(ιq−sq)−1218(52s)[1+∑5j=1|ςj(s)|]ds. | (5.4) |
Let q=2 and εi>0,i=1,2,...,n, satisfying the following inequality.
|D12,120+,ι2ςi(ι)−Fi(ι,ς1(ι),ς2(ι),......,ςn(ι))|≤εi. |
Then, system (3.24) is UH stable with
‖(z1,z2,....,zn)−(ς1,ς2,....,ςn)‖B≤n∑i=1εiciΥi, |
where
Υi=2(22−1)12Γ(12+1)+n∑j≠iκjFiΓ(γj)Γ(12+γj)(22−1)12+γj−γi>0, |
and
ci=1−2118Γ(34)Γ(12+34)(22−1)12=0.7>0. |
We fix i=2 and present the graphical presentations of solutions for the above four cases in Figures 1, 2, 3, and 4 respectively as follows:
In Figure 1, we have presented the graphical presentation of Case I using i=2 to investigate the properties of the system under the Banach weighted spaces and provide insights into the system's response by using the weighted measurable function w(ι)=eι over the interval [0,1]. From Figure 1, the UH stability is obvious and the symmetry in the approximation solution also arises. Moreover, the uniqueness of the solutions here clearly corresponds to the coupled system of two equations. Remember, we use for graphical presentation (1.1) using i=2.
In Figure 2, we have presented the approximate solution for the system using i=2 to investigate the properties of the system under the Banach weighted spaces and provide insights into the system's response by using the weighted measurable function w(ι)=ι over the interval [0,1]. The graphical illustration for the approximate solution demonstrates the symmetry in both solutions for the given weighted measurable function. Here, we have also used (5.2) to illustrate graphs for the approximate solution using i=2.
In Figure 3, the approximate solutions using the weighted measurable function w(ι)=logι in Banach weighted spaces are displayed by selecting the interval [1,2]. We considered again (5.3) using i=2 and demonstrated the approximate solutions to elaborate the behaviors of both graphs. From the graphs, we see that, as logι is considered, we have displayed the approximate solutions over the interval [1,2]. The UH stability can be observed in the graphs and the symmetry also is understandable. There is a little bit of difference in both graphs' behaviors is due to the reason that we do not know the exact solution as the problem is very highly nonlinear.
In Figure 4, we have displayed the approximate solution of (5.4) using i=2 by utilizing the weighted measurable function w(ι)=ιq in Banach weighted spaces. The solutions for the coupled system have been plotted over [1,2]. From the graphs, we see the symmetric behaviors and UH stability. From these discussions, we conclude that symmetry concepts are significant when it comes to natural laws. Further, it is highly interesting that symmetry is the key to nature, but symmetry-breaking processes are mostly responsible for the texture of the universe. There are many different ways that nature's symmetry can be obscured or disturbed. All these discussions about the symmetry in solutions of type as mentioned above integrals will be helpful for those researchers working in physical sciences including quantum physics, mechanics, etc.
In our study, we focused on exploring the existence and uniqueness of symmetric solutions for a system of terminal FDEs in the context of weighted spaces. The specific operator we utilized was the w-Hilfer FD, which is known for its importance in fractional calculus. To examine the existence and uniqueness of symmetric solutions, we employed two fixed-point theorems, the Schauder and Banach fixed-point theorems. These theorems provide the mathematical tools for establishing the existence and uniqueness of solutions in various settings. By applying Schauder and Banach fixed-point theorems to our system, we were able to obtain two key theoretical findings and four symmetry cases. These findings clarify the behavior and properties of symmetric solutions in the considered Banach weighted spaces. By mathematical analysis, we discussed two types of stability in the Ulam sense. The investigation of terminal FDEs using the w-Hilfer FD operator in Banach weighted spaces represents a recent and cutting-edge research direction. We believe this study contributes to the advancement of knowledge in this particular area. Overall, our study not only addresses the challenge of examining the existence and uniqueness of symmetric solutions, but also provides valuable insights into the dynamics and applications of such solutions in fractal mediums. By offering a comprehensive explanation of these processes, our research contributes to the existing literature and expands the understanding of the real-world application of fractional calculus. In future work, we will extend these results with a piecewise fractional operator with the concept of Atangana-Baleanu.
The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.
This work was supported by the Deanship of Scientific Research, Vice Presidency for Graduate Studies and Scientific Research, King Faisal University, Saudi Arabia (grant no. 6057). This study is supported via funding from Prince Sattam bin Abdulaziz University, project number (PSAU/2024/R/1445). The researchers wish to extend their sincere gratitude to the Deanship of Scientific Research at the Islamic University of Madinah for the support provided to the Post-Publishing Program.
The author declares that the research was conducted in the absence of any conflict of interest.
[1] |
A. Thakkar, K. Chaudhari, Fusion in stock market prediction: A decade survey on the necessity, recent developments, and potential future directions, Inform. Fusion, 65 (2021), 95–107. https://doi.org/10.1016/j.inffus.2020.08.019 doi: 10.1016/j.inffus.2020.08.019
![]() |
[2] |
G. Kumar, S. Jain, U. P. Singh, Stock market forecasting using computational intelligence: A survey, Arch. Computat. Methods Eng., 28 (2021), 1069–1101. https://doi.org/10.1007/s11831-020-09413-5 doi: 10.1007/s11831-020-09413-5
![]() |
[3] |
M. Abreu, V. Mendes, Do individual investors trade differently in different financial markets, The Eur. J. Financ., 26 (2020), 1253–1270. https://doi.org/10.1080/1351847X.2019.1709524 doi: 10.1080/1351847X.2019.1709524
![]() |
[4] |
J. F. Ma, J. D. Harstvedt, R. Jaradat, B. Smith, Sustainability driven multi-criteria project portfolio selection under uncertain decision-making environment, Comput. Ind. Eng., 140 (2020), 106236. https://doi.org/10.1016/j.cie.2019.106236 doi: 10.1016/j.cie.2019.106236
![]() |
[5] |
J. W. Goodell, S. Kumar, W. M. Lim, D. Pattnaik, Artificial intelligence and machine learning in finance: Identifying foundations, themes, and research clusters from bibliometric analysis, J. Behav. Exp. Financ., 32 (2021), 100577. https://doi.org/10.1016/j.jbef.2021.100577 doi: 10.1016/j.jbef.2021.100577
![]() |
[6] |
S. R. Das, D. Ostrov, A. Radhakrishnan, D. Srivastav, Dynamic portfolio allocation in goals-based wealth management, Comput. Manag. Sci., 17 (2020), 613–640. https://doi.org/10.1007/s10287-019-00351-7 doi: 10.1007/s10287-019-00351-7
![]() |
[7] |
M. R. Galankashi, F. M. Rafiei, M. Ghezelbash, Portfolio selection: A fuzzy-ANP approach, Financ. Innov., 6 (2020), 17. https://doi.org/10.1186/s40854-020-00175-4 doi: 10.1186/s40854-020-00175-4
![]() |
[8] |
H. Taherdoost, M. Madanchian, Multi-criteria decision making (MCDM) methods and concepts, Encyclopedia, 3 (2023), 77–87. https://doi.org/10.3390/encyclopedia3010006 doi: 10.3390/encyclopedia3010006
![]() |
[9] |
L. A. Zadeh, Fuzzy sets, Information and Control, 8 (1965), 338–353. https://doi.org/10.1016/S0019-9958(65)90241-X doi: 10.1016/S0019-9958(65)90241-X
![]() |
[10] |
K. T. Atanassov, Intuitionistic fuzzy sets, Fuzzy Set. Syst., 20 (1986), 87–96. https://doi.org/10.1016/S0165-0114(86)80034-3 doi: 10.1016/S0165-0114(86)80034-3
![]() |
[11] | I. Despi, D. Opris, E. Yalcin, Generalised Atanassov intuitionistic fuzzy sets, The Fifth International Conference on Information, Process, and Knowledge Management, France, 2013, 51–56. |
[12] |
S. Perçin, Circular supplier selection using interval-valued intuitionistic fuzzy sets, Environ. Dev. Sustain., 24 (2022), 5551–5581. https://doi.org/10.1007/s10668-021-01671-y doi: 10.1007/s10668-021-01671-y
![]() |
[13] |
S. N. M. Kodukulla, V. Sireesha, A novel ranking approach to rank interval-valued trapezoidal intuitionistic fuzzy sets by using the concept of centroids, SN Comput. Sci., 5 (2024), 493. https://doi.org/10.1007/s42979-024-02861-x doi: 10.1007/s42979-024-02861-x
![]() |
[14] |
S. N. M. Kodukulla, V. Sireesha, An extended VIKOR method for interval-valued trapezoidal intuitionistic fuzzy decision making, Mathematical Statistician and Engineering Applications, 71 (2022), 4390–4403. https://doi.org/10.17762/msea.v71i4.1025 doi: 10.17762/msea.v71i4.1025
![]() |
[15] | I. Singuluri, N. R. Shankar, A contemporary transportation perspective for determination of type–1 triangular intuitionistic fuzzy transportation problem, Turk. J. Physiother. R., 32 (2021), 7157–7163. |
[16] |
M. Qiyas, S. Abdullah, Y. Liu, M. Naeem, Multi-criteria decision support systems based on linguistic intuitionistic cubic fuzzy aggregation operators, J. Ambient Intell. Human. Comput., 12 (2021), 8285–8303. https://doi.org/10.1007/s12652-020-02563-1 doi: 10.1007/s12652-020-02563-1
![]() |
[17] |
M. H. Alamoudi, O. A. Bafail, BWM—RAPS approach for evaluating and ranking banking sector companies based on their financial indicators in the Saudi stock market, J. Risk Financial Manag., 15 (2022), 467. https://doi.org/10.3390/jrfm15100467 doi: 10.3390/jrfm15100467
![]() |
[18] |
V. L. G. Nayagam, J. Murugan, Triangular approximation of intuitionistic fuzzy numbers on multi-criteria decision making problem, Soft Comput., 25 (2021), 9887–9914. https://doi.org/10.1007/s00500-020-05346-0 doi: 10.1007/s00500-020-05346-0
![]() |
[19] |
H. Hashemi, P. Ghoddousi, F. Nasirzadeh, Sustainability indicator selection by a novel triangular intuitionistic fuzzy decision-making approach in highway construction projects, Sustainability, 13 (2021), 1477. https://doi.org/10.3390/su13031477 doi: 10.3390/su13031477
![]() |
[20] | F. D. Remadi, H. M. Frikha, The triangular intuitionistic fuzzy extension of the CODAS method for solving multi-criteria group decision making, 2020 International Multi-Conference on: "Organization of Knowledge and Advanced Technologies", Tunis, Tunisia, 2020, 1–6. https://doi.org/10.1109/OCTA49274.2020.9151786 |
[21] |
F. D. Remadi, H. M. Frikha, The triangular intuitionistic fuzzy numbers CODAS method for solving green material selection problem, International Journal of Operational Research, 46 (2023), 398–415. https://doi.org/10.1504/IJOR.2023.129409 doi: 10.1504/IJOR.2023.129409
![]() |
[22] |
M. Saeed, A. Mehmood, A. Anwar, An extension of TOPSIS based on linguistic terms in triangular intuitionistic fuzzy structure, Punjab Univ. J. Math., 53 (2021), 409–424. https://doi.org/10.52280/pujm.2021.530604 doi: 10.52280/pujm.2021.530604
![]() |
[23] | A. Prakash, M. Suresh, Ranking triangular intuitionistic fuzzy numbers: A nagel point approach and applications in multi-criteria decision making, in press. https://doi.org/10.21203/rs.3.rs-3897760/v1 |
[24] |
O. M. Babatunde, J. L. Munda, Y. Hamam, Triangular intuitionistic fuzzy aggregating and ranking function approach for the rating of battery 'End-of-Life' handling alternatives, Energies, 15 (2022), 2248. https://doi.org/10.3390/en15062248 doi: 10.3390/en15062248
![]() |
[25] |
X. L. Geng, Y. Ma, N-Intuitionistic polygonal fuzzy aggregation operators and their application to multi-attribute decision making, IEEE Access, 8 (2020), 162903–162916. https://doi.org/10.1109/ACCESS.2020.3018957 doi: 10.1109/ACCESS.2020.3018957
![]() |
[26] |
S. S. Biswas, B. Alam, M. N. Doja, An algorithm for extracting intuitionistic fuzzy shortest path in a graph, Appl. Comput. Intell. S., 2013 (2013), 970197. https://doi.org/10.1155/2013/970197 doi: 10.1155/2013/970197
![]() |
[27] |
K. Bisht, A. Kumar, Stock portfolio selection hybridizing fuzzy base-criterion method and evidence theory in triangular fuzzy environment, Oper. Res. Forum, 3 (2022), 53. https://doi.org/10.1007/s43069-022-00167-3 doi: 10.1007/s43069-022-00167-3
![]() |
[28] |
H. B. Jiang, B. Q. Hu, A decision-theoretic fuzzy rough set in hesitant fuzzy information systems and its application in multi-attribute decision-making, Inform. Sci., 579 (2021), 103–127. https://doi.org/10.1016/j.ins.2021.07.094 doi: 10.1016/j.ins.2021.07.094
![]() |
[29] |
J. D. Zhou, X. Li, Multi-period mean-semi-entropy portfolio management with transaction costs and bankruptcy control, J. Ambient Intell. Human. Comput., 12 (2021), 705–715. https://doi.org/10.1007/s12652-020-02053-4 doi: 10.1007/s12652-020-02053-4
![]() |
[30] |
M. Narang, M. C. Joshi, K. Bisht, A. Pal, Stock portfolio selection using a new decision-making approach based on the integration of fuzzy CoCoSo with Heronian mean operator, Decision Making: Applications in Management and Engineering, 5 (2022), 90–112. https://doi.org/10.31181/dmame0310022022n doi: 10.31181/dmame0310022022n
![]() |
[31] |
M. Jagtap, P. Karande, The m-polar fuzzy set ELECTRE-I with revised Simos' and AHP weight calculation methods for selection of non-traditional machining processes, Decision Making: Applications in Management and Engineering, 6 (2023), 240–281. https://doi.org/10.31181/dmame060129022023j doi: 10.31181/dmame060129022023j
![]() |
[32] |
A. R. Mishra, P. Rani, F. Cavallaro, A. F. Alrasheedi, Assessment of sustainable wastewater treatment technologies using interval-valued intuitionistic fuzzy distance measure-based MAIRCA method, Facta Univ.-Ser. Mech., 21 (2023), 359–386. https://doi.org/10.22190/FUME230901034M doi: 10.22190/FUME230901034M
![]() |
[33] |
R. Isabels, A. F. Vinodhini, A. Viswanathan, Evaluating and ranking metaverse platforms using intuitionistic trapezoidal fuzzy VIKOR MCDM: Incorporating score and accuracy functions for comprehensive assessment, Decision Making: Applications in Management and Engineering, 7 (2024), 54–78. https://doi.org/10.31181/dmame712024858 doi: 10.31181/dmame712024858
![]() |
[34] |
P. Wang, B. Y. Zhu, Y. Yu, Z. Ali, B. Almohsen, Complex intuitionistic fuzzy DOMBI prioritized aggregation operators and their application for resilient green supplier selection, Facta Univ.-Ser. Mech., 21 (2023), 339–357. https://doi.org/10.22190/FUME230805029W doi: 10.22190/FUME230805029W
![]() |
[35] | S. Sharma, S. Chandra, K. A. Singh, Total productive maintenance using cosine maximization multi attribute method, Ad Alta Interdiscip., 2020 (2020), 110–117. |
[36] |
J. X. Tang, J. D. Cheng, D. L. Xiang, C. B. Hu, Large-difference-scale target detection using a revised Bhattacharyya distance in SAR images, IEEE Geosci. Remote S., 19 (2022), 4506205. https://doi.org/10.1109/LGRS.2022.3161931 doi: 10.1109/LGRS.2022.3161931
![]() |
[37] |
P. F. Zhang, D. X. Wang, Z. Yu, Y. J. Zhang, T. Jiang, T. R. Li, A multi-scale information fusion-based multiple correlations for unsupervised attribute selection, Inform. Fusion, 106 (2024), 102276. https://doi.org/10.1016/j.inffus.2024.102276 doi: 10.1016/j.inffus.2024.102276
![]() |
[38] |
P. F. Zhang, T. R. Li, Z. Yuan, Z. X. Deng, G. Q. Wang, D. X. Wang, et al., A possibilistic information fusion-based unsupervised feature selection method using information quality measures, IEEE T. Fuzzy Syst., 31 (2023), 2975–2988. https://doi.org/10.1109/TFUZZ.2023.3238803 doi: 10.1109/TFUZZ.2023.3238803
![]() |
[39] |
K. Bhattacharya, S. K. De, Decision making under intuitionistic fuzzy metric distances, Annals of Optimization Theory and Practice, 3 (2020), 49–64. https://doi.org/10.22121/aotp.2020.250749.1043 doi: 10.22121/aotp.2020.250749.1043
![]() |
[40] |
A. Bhattacharya, M. Pal, Fuzzy covering problem of fuzzy graphs and its application to investigate the Indian economy in new normal, J. Appl. Math. Comput., 68 (2022), 479–510. https://doi.org/10.1007/s12190-021-01539-4 doi: 10.1007/s12190-021-01539-4
![]() |
[41] |
S. F. Bi, M. Broggi, M. Beer, The role of the Bhattacharyya distance in stochastic model updating, Mech. Syst. Signal Pr., 117 (2019), 437–452. https://doi.org/10.1016/j.ymssp.2018.08.017 doi: 10.1016/j.ymssp.2018.08.017
![]() |
[42] | F. J. Aherne, N. A. Thacker, P. I. Rockett, The Bhattacharyya metric as an absolute similarity measure for frequency coded data, Kybernetika, 34 (1998), 363–368. |
1. | Muhammad Akram, Ayesha Shareef, Ahmad N. Al-Kenani, Pythagorean fuzzy incidence graphs with application in one-way toll road network, 2024, 9, 2364-4966, 10.1007/s41066-024-00455-9 | |
2. | Muhammad Akram, Uzma Ahmad, Mohammed M. Ali Al-Shamiri, Ayesha Shareef, Algorithms for computing Pythagorean fuzzy average edge connectivity of Pythagorean fuzzy graphs, 2024, 1598-5865, 10.1007/s12190-023-01970-9 |