Loading [MathJax]/jax/output/SVG/jax.js
Research article Special Issues

Stabilization of nonlinear hybrid stochastic time-delay neural networks with Lévy noise using discrete-time feedback control

  • This paper aims to formulate a class of nonlinear hybrid stochastic time-delay neural networks (STDNNs) with Lévy noise. Specifically, the coefficients of networks grow polynomially instead of linearly, and the time delay of given neural networks is non-differentiable. In many practical situations, nonlinear hybrid STDNNs with Lévy noise are unstable. Hence, this paper uses feedback control based on discrete-time state and mode observations to stabilize the considered nonlinear hybrid STDNNs with Lévy noise. Then, we establish stabilization criteria of H stability, asymptotic stability, and exponential stability for the controlled nonlinear hybrid STDNNs with Lévy noise. Finally, a numerical example illustrating the usefulness of theoretical results is provided.

    Citation: Tian Xu, Ailong Wu. Stabilization of nonlinear hybrid stochastic time-delay neural networks with Lévy noise using discrete-time feedback control[J]. AIMS Mathematics, 2024, 9(10): 27080-27101. doi: 10.3934/math.20241317

    Related Papers:

    [1] Fugeng Zeng, Peng Shi, Min Jiang . Global existence and finite time blow-up for a class of fractional p-Laplacian Kirchhoff type equations with logarithmic nonlinearity. AIMS Mathematics, 2021, 6(3): 2559-2578. doi: 10.3934/math.2021155
    [2] Sarra Toualbia, Abderrahmane Zaraï, Salah Boulaaras . Decay estimate and non-extinction of solutions of p-Laplacian nonlocal heat equations. AIMS Mathematics, 2020, 5(3): 1663-1679. doi: 10.3934/math.2020112
    [3] Hao Zhang, Zejian Cui, Xiangyu Xiao . Decay estimate and blow-up for a fourth order parabolic equation modeling epitaxial thin film growth. AIMS Mathematics, 2023, 8(5): 11297-11311. doi: 10.3934/math.2023572
    [4] Zayd Hajjej, Sun-Hye Park . Asymptotic stability of a quasi-linear viscoelastic Kirchhoff plate equation with logarithmic source and time delay. AIMS Mathematics, 2023, 8(10): 24087-24115. doi: 10.3934/math.20231228
    [5] Salim A. Messaoudi, Mohammad M. Al-Gharabli, Adel M. Al-Mahdi, Mohammed A. Al-Osta . A coupled system of Laplacian and bi-Laplacian equations with nonlinear dampings and source terms of variable-exponents nonlinearities: Existence, uniqueness, blow-up and a large-time asymptotic behavior. AIMS Mathematics, 2023, 8(4): 7933-7966. doi: 10.3934/math.2023400
    [6] Zhiqiang Li . The finite time blow-up for Caputo-Hadamard fractional diffusion equation involving nonlinear memory. AIMS Mathematics, 2022, 7(7): 12913-12934. doi: 10.3934/math.2022715
    [7] Abdelbaki Choucha, Salah Boulaaras, Asma Alharbi . Global existence and asymptotic behavior for a viscoelastic Kirchhoff equation with a logarithmic nonlinearity, distributed delay and Balakrishnan-Taylor damping terms. AIMS Mathematics, 2022, 7(3): 4517-4539. doi: 10.3934/math.2022252
    [8] Mohammad Kafini, Shadi Al-Omari . Local existence and lower bound of blow-up time to a Cauchy problem of a coupled nonlinear wave equations. AIMS Mathematics, 2021, 6(8): 9059-9074. doi: 10.3934/math.2021526
    [9] Khaled Zennir, Abderrahmane Beniani, Belhadji Bochra, Loay Alkhalifa . Destruction of solutions for class of wave p(x)bi-Laplace equation with nonlinear dissipation. AIMS Mathematics, 2023, 8(1): 285-294. doi: 10.3934/math.2023013
    [10] Adel M. Al-Mahdi . The coupling system of Kirchhoff and Euler-Bernoulli plates with logarithmic source terms: Strong damping versus weak damping of variable-exponent type. AIMS Mathematics, 2023, 8(11): 27439-27459. doi: 10.3934/math.20231404
  • This paper aims to formulate a class of nonlinear hybrid stochastic time-delay neural networks (STDNNs) with Lévy noise. Specifically, the coefficients of networks grow polynomially instead of linearly, and the time delay of given neural networks is non-differentiable. In many practical situations, nonlinear hybrid STDNNs with Lévy noise are unstable. Hence, this paper uses feedback control based on discrete-time state and mode observations to stabilize the considered nonlinear hybrid STDNNs with Lévy noise. Then, we establish stabilization criteria of H stability, asymptotic stability, and exponential stability for the controlled nonlinear hybrid STDNNs with Lévy noise. Finally, a numerical example illustrating the usefulness of theoretical results is provided.



    If the image of a point x under two single-valued mappings is x itself, then x is said to be a fixed point of these mappings. Banach [7] proved a meaningful result for contraction mappings. Due to its significance, several authors, like Acar et al. [3], Altun et al. [5], Aslantas et al. [6], Sahin et al. [27], Hussain et al. [17], Hammad et al. [14,15,16] and Ceng et al. [8,9,10,11] presented many related useful applications in fixed point theory. In [23,31], the authors showed a new iterative scheme for the solution of nonlinear mixed Volterra Fredholm type fractional delay integro-differential equations of different orders. Chistyakov [13] introduced the notion of a modular metric space. Mongkolkeha et al. [21] established some results in modular metric spaces for contraction mappings. Chaipunya et al. [12], Abdou et al. [2] and Alfuraidan et al. [4] showed fixed point results for multivalued mappings in modular metric spaces. Abdou et al. [1] proved fixed point theorems of pointwise contractions in modular metric spaces. Hussain et al. [19] discussed some fixed point theorems for generalized F-contractions in fuzzy metric and modular metric spaces. Later, Padcharoen et al. [22] introduced the concept of α-type F-contractions in modular metric spaces and showed fixed point and periodic point results for such a contraction. Recently, Rasham et al. [26] introduced a modular-like metric space and proved results for families of mappings in such spaces. In this research work, we prove existence of fixed point results for a hybrid pair of multivalued maps fulfilling generalized rational type F-contractions, by using a weaker class of strictly increasing mappings F rather than the class of mappings introduced by Wardowski [30].

    Let us state the following preliminary concepts.

    Definition 1.1. [26] Let B be a non-empty set. A function υ:(0,)×B×B[0,) is said to be a modular-like metric on B, if for each e,i,oB and υ(a,i,o)=υa(i,o), the following hold:

    (i) υa(i,o)=υa(o,i) for all a>0;

    (ii) υa(i,o)=0 for all a>0 implies i=o;

    (iii) υl+n(i,o)υl(i,e)+υn(e,o) for all l,n>0.

    The pair (B,υ) is said to be a modular-like metric space. If we change (ii) by "υl(i,o)=0 for each l>0 iff i=o", then (B,υ) becomes a modular metric space. While, by changing (ii) with "υl(i,o)=0 for some l>0, such that i=o", we obtain a regular modular-like metric space. For sB and ε>0, ¯Cυl(s,ε)={tB:|υl(s,t)υl(t,t)|ε} is a closed ball in (B,υ).

    Example 1.2. Let B=[0,)×[0,). Define υ:(0,)×B×B[0,) as

    (i)υ(a,(e,p),(i,o))=e+p+i+oa,(ii)υ(a,(e,p),(i,o))=max{e,p,i,o}a.

    The functions given in (i) and (ii) are examples of a modular-like metric on B.

    Definition 1.3. [26] Let (B,υ) be a modular-like metric space.

    (i) A sequence (an)nN in B is said to be υ-convergent to a point aB for some l>0 if limn+υl(an,a)=υl(a,a).

    (ii) A sequence (an)nN in B is said to be an υ-Cauchy sequence for some l>0 if limn,mυl(am,an) exists and is finite.

    (iii) B is called υ-complete if each υ-Cauchy sequence (an)nN in B is υ-convergent to some aB, that is,

    limn+υl(an,a)=υl(a,a).

    (iv) If every sequence has a convergent subsequence, then B is called compact.

    Definition 1.4. [26] Let (B,υ) be a modular-like metric space and UB. An element p0 in U verifying

    υl(s,U)=infp0Uυl(s,p0)

    is called a best approximation in U for sB. If each sB possesses a best approximation in U, then U is called a proximinal set.

    From now on, let P(B) represent the set of proximinal compact subsets in B.

    Example 1.5. Let B=[0,) and υl(s,r)=1w(s+r) with w>0. Take U=[7,8]. Then for any mB,

    υl(m,U)=υl(m,[7,8])=infn[7,8]υl(m,n)=υl(m,7).

    So 7 is a best approximation in U for any mB. Moreover, [7,8] is a proximinal set.

    Definition 1.6. [26] The mapping Hυl:P(B)×P(B)[0,), given by

    Hυl(X,Y)=max{supσXυl(σ,Y),supςYυl(ς,X)},

    is known as an υl- Hausdorff metric. Note that (P(B),Hυl) is named as an υl- Hausdorff metric space.

    Example 1.7. Let B=[0,) and υl(θ,ϑ)=1l(θ+ϑ) with l>0. Taking W=[5,6] and Q=[9,10] we get Hυl(W,Q)=15l.

    Definition 1.8. [26] Let (X,υ) be a modular-like metric space. υ is said to satisfy the M-condition if limn,mυp(xn,xm)=0, where pN implies limn,mυl(xn,xm)=0, for some l>0.

    Definition 1.9. [28] Let CΦ, Y:CP(C) be a multivalued mapping, EC and α:C×C[0,+) be a function. Then Y is said to be α-admissible on E if α(Ye,Yz)=inf{α(l,m):lYe,mYz}1, whenever α(e,z)1 for all e, z E.

    Definition 1.10. [29] Let BΦ, Y:BP(B) be a multi-valued mapping, RB and α:B×B[0,) be a function. Then Y is said to be α-dominated on R if for all vR, α(v,Yv)=inf{α(v,j):jYv}1.

    Definition 1.11. [30] Let (C,d) be a metric space. A self mapping H:CC is said to be a Q-contraction if for each g,kC, there is τ>0 such that d(Ca,Cg)>0 implies

    τ+Q(d(Ca,Cg))Q(d(a,g)),

    where Q:(0,)R satisfies the following:

    (F1) For any k(0,1), limσ0+σkQ(σ)=0;

    (F2) For each u,v>0 such that u<v, Q(u)<Q(v);

    (F3) limn+σn =0 if and only if limn+Q(σn)= for every positive sequence {σn}n=1.

    Let ϝ denote the set of mappings such that (F1)–(F3) hold.

    Lemma 1.12. [26] Let (£,υ) be amodular-like metric space. Let (P(£),Hυl) be aHausdorff υlmetric-like space. Then, for all bU and foreach U,YP(£), there is baY such that Hυl(U,Y)υl(a,ba).

    Example 1.13. [24] Let W=R. Consider α:W×W[0,) as

    α(s,r)={1ifs>r14ifsr.

    Define L,N:WP(W) by

    Ls=[4+s,3+s]andNr=[2+r,1+r].

    The α-dominated property for L and N holds. Note that L and N are not α-admissible.

    Let (£,υ) be a modular-like metric space, δ0£, and R,C:£ P(£) be two multifunctions on £. For δ1Rδ0 with υ1(δ0,Rδ0)=υ1(δ0,δ1), take δ2Cδ1 such that υ1(δ1,Cδ1)=υ1(δ1,δ2). Choose δ3Rδ2 such that υ1(δ2,Rδ2)=υ1(δ2,δ3). In this way, we get a sequence {CR(δn)} in £, where

    δ2n+1Rδ2n,δ2n+2Cδ2n+1,

    for all nN{0}. Note that υ1(δ2n,Rδ2n)=υ1(δ2n,δ2n+1) and υ1(δ2n+1,Cδ2n+1)=υ1(δ2n+1,δ2n+2). {CR(δn)} is said to be a sequence in £ generated by δ0. If R=C, then we denote {£R(δn)} instead of {CR(δn)}.

    Theorem 2.1. Let (£,υ) be a completemodular-like metric space. Suppose that υ is regular and verifiesthe M-condition. Let δ0 £, α:£×£[0,) and R,C:£ P(£) be α-dominatedmultifunctions on £. Assume there are τ>0 and Qϝsuch that

    τ+Q(Hυ1(Rt,Cδ))Q(max{υ1(t,δ),υ1(t,Rt),υ2(t,Cδ)2,υ1(t,Rt).υ1(δ,Cδ)1+υ1(t,δ)}) (2.1)

    where t,δ{CR(δn)}, α(t,δ)1 or α(δ,t)1, and Hυ1(Rt,Cδ)>0. Then the sequence {CR(δn)} generated by δ0 convergesto e£ and for each nN, α(δn,δn+1)1. Furthermore, if e satisfies (2.1), α(δn,e)1 and α(e,δn)1 forall integers n0, then R and C have a common fixed point e in £.

    Proof. Consider a sequence {CR(δn)}. Obviously, δn£ for each integer n0. If j is odd, then j=2ˊı+1 for some ˊıN. By definition of α-dominated mappings, one has α(δ2ˊı,Rδ2ˊı)1 and α(δ2ˊı+1,Cδ2ˊı+1)1. Since α(δ2ˊı,Rδ2ˊı)1, one gets inf{α(δ2ˊı,b):bRδ2ˊı}1. Also, δ2ˊı+1Rδ2ˊı and so α(δ2ˊı,δ2ˊı+1)1. Moreover, δ2ˊı+2Cδ2ˊı+1 and so α(δ2ˊı+1,δ2ˊı+2)1. In view of Lemma 1.12, we have

    τ+Q(υ1(δ2ˊı+1,δ2ˊı+2))τ+Q(Hυ1(Rδ2ˊı,Cδ2ˊı+1))Q(max{υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı,Rδ2ˊı),υ2(δ2ˊı,Cδ2ˊı+1)2,υ1(δ2ˊı,Rδ2ˊı).υ1(δ2ˊı+1,Cδ2ˊı+1)1+υ1(δ2ˊı,δ2ˊı+1)})Q(max{υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı,δ2ˊı+1)+υ1(δ2ˊı+1,δ2ˊı+2)2,υ1(δ2ˊı,δ2ˊı+1).υ1(δ2ˊı+1,δ2ˊı+2)1+υ1(δ2ˊı,δ2ˊı+1)})Q(max{υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı+1,δ2ˊı+2)}).

    This implies

    τ+Q(υ1(δ2ˊı+1,δ2ˊı+2))Q(max{υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı+1,δ2ˊı+2)}). (2.2)

    Now, if

    max{υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı+1,δ2ˊı+2)}=υ1(δ2ˊı+1,δ2ˊı+2),

    then from (2.2), we have

    Q(υ1(δ2ˊı+1,δ2ˊı+2))Q(υ1(δ2ˊı+1,δ2ˊı+2))τ,

    which is a contradiction. Therefore,

    max{υ1(δ2ˊı,δ2ˊı+1),υ1(δ2ˊı+1,δ2ˊı+2)}=υ1(δ2ˊı,δ2ˊı+1)

    for all ˊı0. Hence, from (2.2), we have

    Q(υ1(δ2ˊı+1,δ2ˊı+2))Q(υ1(δ2ˊı,δ2ˊı+1))τ. (2.3)

    Similarly, we have

    Q(υ1(δ2ˊı,δ2ˊı+1))Q(υ1(δ2ˊı1,δ2ˊı))τ (2.4)

    for all ˊı0. By (2.3) and (2.4), we have

    Q(υ1(δ2ˊı+1,δ2ˊı+2))Q(υ1(δ2ˊı1,δ2ˊı))2τ.

    Repeating these steps, we get

    Q(υ1(δ2ˊı+1,δ2ˊı+2))Q(υ1(δ0,δ1))(2ˊı+1)τ. (2.5)

    Similarly, we have

    Q(υ1(δ2ˊı,δ2ˊı+1))Q(υ1(δ0,δ1))2ˊıτ. (2.6)

    By (2.5) and (2.6), we obtain

    Q(υ1(δn,δn+1))Q(υ1(δ0,δ1))nτ. (2.7)

    Letting n in (2.7), one obtains

    limnQ(υ1(δn,δn+1))=.

    Since Qϝ,

    limnυ1(δn,δn+1)=0. (2.8)

    Due to (F1) of ϝ, there is k(0,1) such that

    limn(υ1(δn,δn+1))k(Q(υ1(δn,δn+1))=0. (2.9)

    By (2.7), for all nN, we obtain

    (υ1(δn,δn+1))k(Q(υ1(δn,δn+1))Q(υ1(δ0,δ1))(υ1(δn,δn+1))knτ0. (2.10)

    Using (2.8), (2.9) and taking n in (2.10), we have

    limnn(υ1(δn,δn+1))k=0. (2.11)

    By (2.11), there is n1N such that n(υ1(δn,δn+1))k1 for all nn1, or

    υ1(δn,δn+1)1n1kforallnn1.

    Letting p>0 and m=n+p> n>n1, we get

    υp(δn,δm)υ1(δn,δn+1)+υ1(δn+1,δn+2)++υ1(δm,δm+1)j=n1j1k.

    Since k(0,1), 1k>1 and so the series j=11j1k converges. Thus,

    limn,mυp(δn,δm)=0.

    Since υ satisfies the M-condition, we have

    limn,mυ1(δn,δm)=0. (2.12)

    Hence {CR(δn)} is Cauchy in the regular complete modular-like metric space (£,υ) and so there is e£ such that {CR(δn)}e as n and thus

    limnυ1(δn,e)=0. (2.13)

    Now, by Lemma 1.12, one obtains

    τ+Q(υ1(δ2n+1,Ce)τ+Q(Hυ1(Rδ2n,Ce)). (2.14)

    Now, there exists δ2n+1Rδ2n such that υ1(δ2n,Rδ2n)=υ1(δ2n,δ2n+1). From assumption, α(δn,e)1. Assume that υ1(e,Ce)>0. Then there is an integer p>0 such that υ1(δ2n+1,Ce)>0 for np. Now, if Hυ1(Rδ2n,Ce)>0, then by (2.1), we have

    τ+Q(υ1(δ2n+1,Ce))Q(max{υ1(δ2n,e),υ1(δ2n,e),υ1(δ2n,δ2n+1)+υ1(δ2n+1,Ce)2,υ1(δ2n,Rδ2n).υ1(Q,Ce)1+υ1(δ2n,e)}).

    Letting n and using (2.13), we get

    τ+Q(υ1(e,Ce))Q(υ1(e,Ce)).

    Since Q is strictly increasing, (2.14) implies

    υ1(e,Ce)<υ1(e,Ce).

    This is a contradiction. Hence υ1(e,Ce)=0 and so eCe.

    Similarly, we can show that υ1(e,Re)=0, that is, eRe. Hence e is a common fixed point of both mappings R and C in £.

    Corollary 2.2. Let (£,υ) be a completemodular-like metric space. Suppose that υ is regular and verifiesthe M-condition. Let α:£×£[0,) and R,C:£ P(£)be α-dominated multifunctions on £. Assume thereare τ>0 and Qϝ such that

    τ+Q(Hυ1(Rt,Cδ))Q(max{υ1(t,δ),υ1(t,Rt),υ2(t,Cδ)2,υ1(t,Rt).υ1(δ,Cδ)1+υ1(t,δ)}),

    where t,δ£, α(t,δ)1 or α(δ,t)1, and Hυ1(Rt,Cδ)>0. Then there exists a sequence {δn} in £ converging to e£ and for each nN, α(δn,δn+1)1. Also, if α(δn,e)1 and α(e,δn)1 for all integers n0, then R and C have a common fixed point e in £.

    Example 2.3. Let £=R+{0}. Take υ2(r,m)=r+m and υ1(e,t)=12(e+t) for all e,t£. Define R,C:£ P(£) by

    Rv={[v3,2v3]ifv{1,13,112,136,1144,}[7v,10v]ifv{1,13,112,136,1144,}

    and

    Cv={[v4,3v4]ifv{1,13,112,136,1144,}[5v,13v]ifv{1,13,112,136,1144,}.

    Suppose that v0=1. Then υ1(v0,Rv0)=υ1(1,R1)=υ1(1,13) and so v1=13. Now, υ1(v1,Cv1)=υ1(13,C13)=υ1(13,112) and thus v2=112. Now, υ1(v2,Rv2)=υ1(112,R112)=υ1(112,136) and so v3=136. Continuing in this way, we have {CR(vn)}={1,13,112,136,1144,}. Define α:£×£[0,) as

    α(r,t)={1     ifr>t12otherwise.

    Let v,y{CR(vn)} with α(v,y)1. Then

    Hυ1(Rv,Cy)=max{supaRvυ1(a,Cy),supbCyυ1(Rv,b)}=max{υ1(2v3,[y4,3y4]),υ1([v3,2v3],3y4)}=max{υ1(2v3,y4),υ1(v3,3y4)}=max{2v3+y4,v3+3y4}.

    Also,

    max{υ1(v,y),υ1(v,Rv),υ2(v,Cy)2,υ1(v,Rv).υ1(y,Cy)1+υ1(v,y)}=max{v+y,v+v3,14(v+y4),(v+v3).(y+y4)1+v+y}.

    If Q(t)=lnt and τ=ln(1.2), then we have

    τ+Q(Hυ1(Rv,Cy))Q(max{υ1(v,y),υ1(v,Rv),υ2(v,Cy)2,υ1(v,Rv).υ1(y,Cy)1+υ1(v,y)}).

    Hence all the conditions in Theorem 2.1 hold and so R and C possess a common fixed point.

    Note that

    Rv={[v3,2v3]ifv{1,13,112,136,1144,}[7v,10v]ifv{1,13,112,136,1144,}

    and

    Cv={[v4,3v4]ifv{1,13,112,136,1144,}[5v,13v]ifv{1,13,112,136,1144,}.

    If v=2 and y=3, then we have

    Hυ1(R2,C3)=max{supaR2υ1(a,C3),supbC3υ1(R2,b)}=max[{supa[14,20]υ1(a,[15,39]),supb[15,39]υ1([14,20],b)}]=max[{supa[14,20]υ1(a,15),supb[15,39]υ1(14,b)}]=max{υ1(20,15),υ1(14,39)}=max{20+15,14+39}=53.

    Also

    max{υ1(v,y),υ1(v,Rv),υ2(v,Cy)2,υ1(v,Rv).υ1(y,Cy)1+υ1(v,y)}=max{υ1(2,3),υ1(2,[14,20]),υ2(2,[15,39])2,υ1(2,[14,20]).υ1(3,[15,39])1+υ1(2,3)}=max{5,16,174,(16)(18)6}=48.

    Now,

    ln(1.2)+ln(53)>ln(48).

    This implies that

    τ+F(Hυ1(R2,C3)>F(υ1(2,3)).

    So the condition (2.1) does not hold on the whole space. Hence Corollary 2.2 and the other existing results in modular metric spaces cannot be applied to ensure the existence of a common fixed point. However, Theorem 2.1 is valid here.

    Taking R=C in Theorem 2.1, we may state the following corollary.

    Corollary 2.4. Let (£,υ) be a completemodular-like metric space. Suppose υ is regular and the M-condition holds. Let δ0£, α:£×£[0,) and R:£ P(£) be a α-dominated set-valuedfunction on £. Assume there are τ>0 and Qϝ suchthat

    τ+Q(Hυl(Rt,Rδ))Q(max{υ1(t,δ),υ1(t,Rt),υ2(t,Rδ),υ1(t,Rt).υ1(δ,Rδ)1+υ1(t,δ)}), (2.15)

    where t,δ{£R(δn)}, α(t,δ)1, and Hυ1(Rt,Rδ)>0. Then, the sequence {£R(δn)} generated by δ0 converges to e£and for each integer n0, α(δn,δn+1)1.Also, if e satisfies (2.15) and either α(δn,e)1 or α(e,δn)1 for all integers n0, then R has afixed point e in £.

    Jachymski [20] initiated the graph concept in fixed point theory. Hussain et al. [18] gave new results for graphic contractions. Recently, Younis et al. [32] discussed a significant result on the graphical structure of extended b-metric spaces and Shoaib et al. [29] established some results on graph dominated set-valued mappings in the setting of b-metric like spaces. Further results on graph theory can be seen in [24,25,28].

    Definition 3.1. [29] Let A be a non-empty set and Υ=(V(Υ),L(Υ)) be a graph with V(Υ)=A. A mapping P from A into P(A) is said to be multi-graph dominated on A if for each ıA, we have (ı,ȷ)L(Υ), where ȷPa.

    Theorem 3.2. Let (U,υ) be a complete modular-likemetric space endowed with a graph Υ and δ0R satisfying the following:

    (i) R and C are multi-graph dominated functions on {CR(δn)};

    (ii) There are τ>0 and Qϝ such that

    τ+Q(Hυ1(Rw,Ch))Q(max{υ1(w,h),υ1(w,Rw),υ2(w,Ch)2,υ1(w,Rw).υ1(h,Ch)1+υ1(w,h)}), (3.1)

    where w,h{CR(δn)}, (w,h)L(Υ)or (h,w)L(Υ), and Hυ1(Rw,Ch)>0.Suppose that the regularity of R and the M-conditionare verified. Then (δn,δn+1)L(Υ)and {CR(δn)}δ. Also, if δ satisfies (3.1), (δn,δ)L(Υ) and (δ,δn)L(Υ)for all integers n0, then R and C have a common fixed point in U.

    Proof. Define α:U×U[0,) as α(w,h)=1 if wU and (w,h)L(Υ), and α(w,h)=0, otherwise. The graph domination on U yields that (w,h)L(Υ) for all hRw and (w,h)L(Υ) for each hCw. So α(w,h)=1 for all hRw and α(w,h)=1 for each hCw. Thus inf{α(w,h):hRw}=1 and inf{α(w,h):hCw}=1. Hence α(w,Rw)=1 and α(w,Cw)=1 for any wR. So R and C are α-dominated on U. Furthermore,

    τ+Q(Hυ1(Rw,Ch))Q(max{υ1(w,h),υ1(w,Rw),υ2(w,Ch)2,υ1(w,Rw).υ1(h,Ch)1+υ1(w,h)}),

    where w,hU{CR(δn)}, α(w,h)1 and Hυ1(Rw,Ch)>0. Also, (ii) is fulfilled. Due to Theorem 2.1, {CR(δn)} is a sequence in U and {CR(δn)}δU. Here, δn,δU and either (δn,δ)L(Υ) or (δ,δn)L(Υ) yields that either α(δn,δ)1 or α(δ,δn)1. So all the hypotheses of Theorem 2.1 hold. Thus δ is a common fixed point of R and C in U and υ1(δ,δ)=0.

    In this section, some corollaries related to single-valued mappings in modular-like metric space are derived. Let (£,υ) be a modular-like metric space, δ0£ and R,C:££ be a pair of mappings. Let δ1=Rδ0, δ2=Cδ1, δ3=Rδ2. Consider a sequence {δn} in £ such that δ2n+1=Rδ2n and δ2n+2=Cδ2n+1, for integers n0. We represent this type of iteration by {CR(δn)}. {CR(δn)} is a sequence in £ generated by δ0. If R=C, then we use {£R(δn)} instead of {CR(δn)}.

    Theorem 4.1. Let (£,υ) be a completemodular-like metric space. Suppose that the regularity of υ andthe M-condition hold. Take r>0, δ0£, α:£×£[0,)and let R,C:££ be α-dominatedmultifunctions on £. Then there are τ>0 and Qϝsuch that

    τ+Q(υ1(Rt,Cδ))Q(max{υ1(t,δ),υ1(t,Rt),υ2(t,Cδ)2,υ1(t,Rt).υ1(δ,Cδ)1+υ1(t,δ)}), (4.1)

    where t,δ{CR(δn)}, α(t,δ)1, or α(δ,t)1, and υ1(Rt,Cδ)>0. Then α(δn,δn+1)1 for all integers n0and {CR(δn)}h£. Also, if h verifies(4.1), α(δn,h)1 and α(h,δn)1 forall integers n0, then R and C admit a common fixed point h in £.

    Proof. The proof is similar to the proof of Theorem 2.1.

    Letting R=C in Theorem 4.1, we have the following corollary.

    Corollary 4.2. Let (£,υ) be a complete modularlike metric space. Suppose that the regularity of υ and the M-condition hold. Choose δ0£, α:£×£[0,) and let R:££ be a single-valued function on £.Then there are τ>0 and Qϝ such that

    τ+Q(υ1(Rt,Rδ))Q(max{υ1(t,δ),υ1(t,Rt),υ2(t,Rδ)2,υ1(t,Rt).υ1(δ,Rδ)1+υ1(t,δ)}), (4.2)

    where t,δ{£R(δn)}, α(t,δ)1, or α(δ,t)1, and υ1(Rt,Rδ)>0.Then α(δn,δn+1)1 for all integers n0and {δn}h£. Also, if (4.2) holds for h, α(δn,h)1 and α(h,δn)1 forall integers n0, then R has a fixed point h.

    In this section, we apply our work to solve integral equations.

    Theorem 5.1. Let (£,υ) be a completemodular-like metric space. Suppose that the regularity of υ andthe M-condition hold. Take r>0, δ0£ and let R,C:££ be α-dominatedmultifunctions on £. Then there are τ>0and Qϝ such that

    τ+Q(υ1(Rt,Cδ))Q(max{υ1(t,δ),υ1(t,Rt),υ2(t,Cδ)2,υ1(t,Rt).υ1(δ,Cδ)1+υ1(t,δ)}), (5.1)

    where t,δ{CR(δn)}, and υ1(Rt,Cδ)>0.Then {CR(δn)}f£. Also, if f verifies (5.1), then R and C admit a unique common fixed point f in £.

    Let W=C([0,1],R+) be the family of continuous functions defined on [0,1]. The following are two integral equations:

    u(e)=e0H(e,f,u(f))df, (5.2)
    c(e)=e0G(e,f,c(f))df (5.3)

    for all e[0,1], where H,G:[0,1]×[0,1]×WR. For δC([0,1],R+), define supremum norm as δη=sups[0,1]{|δ(s)|eτs}, and take τ>0 arbitrarily. For all c,wC([0,1],R+), define

    υ1(δ,w)=12sups[0,1]{|δ(s)+w(s)|eτs}=12δ+wτ.

    It is clear that (C([0,1],R+),dτ) is a complete modular-like metric space. So we have the following result.

    Theorem 5.2. Suppose that

    (i) H,G:[0,1]×[0,1]×C([0,1],R+)R;

    (ii) Define

    (Ru)(e)=e0H(e,f,u(f))df,(Cδ)(e)=e0G(e,f,δ(f))df.

    Assume that there is τ>0 such that

    |H(e,f,u)+G(e,f,δ)|τM(u,δ)τM(u,δ)+1

    for all e,f[0,1] and u,δC([0,1],R+), where

    M(u,δ)=max(12{u+δτ,u+Ruτ,u+Ruτ+δ+Cδτ2,u+Ruτ.δ+Cδτ1+u+δτ}).

    Then (5.2) and (5.3) possess a unique solution.

    Proof. By (ⅱ),

    |Ru+Cδ|=e0|H(e,f,u)+G(e,f,δ)|dfe0τM(u,δ)τM(u,δ)+1eτfdfτM(u,δ)τM(u,δ)+1e0eτfdfM(u,δ)τM(u,δ)+1eτe.

    This implies

    |Ru+Cδ|eτeM(u,δ)τM(u,δ)+1,
    Ru+CδτM(u,δ)τM(u,δ)+1,
    τM(u,δ)+1M(u,δ)1Ru+Cδτ,
    τ+1M(u,δ)1Ru+Cδτ.

    Thus

    τ1Ru(e)+Cδ(e)τ1M(u,δ).

    All the conditions of Theorem 5.1 hold for Q(f)=1f for f >0 and υ1(f,δ)=12f+δτ. Hence both the integral Eqs (5.2) and (5.3) admit a unique common solution.

    In this article, we have achieved some new results for a pair of set-valued mappings verifying a generalized rational Wardowski type contraction. Dominated mappings are applied to obtain some fixed point theorems. Applications on integral equations and graph theory are given. Moreover, we investigate our results in a more better new framework. New results in ordered spaces, modular metric space, dislocated metric space, partial metric space, b-metric space and metric space can be obtained as corollaries of our results. One can further extend our results to fuzzy mappings, bipolar fuzzy mappings and fuzzy neutrosophic soft mappings.

    The authors declare that we have no conflict of interest.



    [1] S. Blythe, X. R. Mao, X. X. Liao, Stability of stochastic delay neural networks, J Franklin Inst., 338 (2001), 481–495. https://doi.org/10.1016/S0016-0032(01)00016-3 doi: 10.1016/S0016-0032(01)00016-3
    [2] Y. Chen, W. X. Zheng, Stability analysis of time-delay neural networks subject to stochastic perturbations, IEEE Trans. Cybern., 43 (2013), 2122–2134. https://doi.org/10.1109/TCYB.2013.2240451 doi: 10.1109/TCYB.2013.2240451
    [3] Z. D. Wang, H. S. Shu, J. A. Fang, X. H. Liu, Robust stability for stochastic Hopfield neural networks with time delays, Nonlinear Anal.: Real World Appl., 7 (2006), 1119–1128. https://doi.org/10.1016/j.nonrwa.2005.10.004 doi: 10.1016/j.nonrwa.2005.10.004
    [4] P. F. Wang, Q. J. He, H. Su, Stabilization of discrete-time stochastic delayed neural networks by intermittent control, IEEE Trans. Cybern., 53 (2021), 2017–2027. https://doi.org/10.1109/TCYB.2021.3108574 doi: 10.1109/TCYB.2021.3108574
    [5] D. N. Lu, D. B. Tong, Q. Y. Chen, W. N. Zhou, J. Zhou, S. G. Shen, Exponential synchronization of stochastic neural networks with time-varying delays and Lévy noises via event-triggered control, Neural Process. Lett., 53 (2021), 2175–2196. https://doi.org/10.1007/s11063-021-10509-7 doi: 10.1007/s11063-021-10509-7
    [6] C. Fei, W. Y. Fei, X. R. Mao, L. T. Yan, Delay-dependent asymptotic stability of highly nonlinear stochastic differential delay equations driven by G-Brownian motion, J. Franklin Inst., 359 (2022), 4366–4392. https://doi.org/10.1016/j.jfranklin.2022.03.027 doi: 10.1016/j.jfranklin.2022.03.027
    [7] Q. He, Y. Ren, Finite-time stabilisation issue for a class of highly nonlinear stochastic coupled systems, Int. J. Control, 2023. https://doi.org/10.1080/00207179.2023.2263580
    [8] A. L. Wu, H. Yu, Z. G. Zeng, Variable-delay feedback control for stabilisation of highly nonlinear hybrid stochastic neural networks with time-varying delays, Int. J. Control, 97 (2024), 744–755. https://doi.org/10.1080/00207179.2023.2168878 doi: 10.1080/00207179.2023.2168878
    [9] E. Fridman, Introduction to time-delay systems: analysis and control, Birkh¨auser: Cham, 2014. https://doi.org/10.1007/978-3-319-09393-2
    [10] H. Huang, D. W. C. Ho, Y. Z. Qu, Robust stability of stochastic delayed additive neural networks with Markovian switching, Neural Netw., 20 (2007), 799–809. https://doi.org/10.1016/j.neunet.2007.07.003 doi: 10.1016/j.neunet.2007.07.003
    [11] L. C. Feng, J. D. Cao, L. Liu, Stability analysis in a class of Markov switched stochastic Hopfield neural networks, Neural Process. Lett., 50 (2019), 413–430. https://doi.org/10.1007/s11063-018-9912-7 doi: 10.1007/s11063-018-9912-7
    [12] C. D. Zheng, Q. H. Shan, H. G. Zhang, Z. S. Wang, On stabilization of stochastic Cohen-Grossberg neural networks with mode-dependent mixed time-delays and Markovian switching, IEEE Trans. Neural Networks Learn. Syst., 24 (2013), 800–811. https://doi.org/10.1109/TNNLS.2013.2244613 doi: 10.1109/TNNLS.2013.2244613
    [13] A. L. Wu, H. Yu, Z. G. Zeng, Stabilization of highly nonlinear hybrid neutral stochastic neural networks with time-varying delays by variable-delay feedback control, Syst. Control Lett., 172 (2023), 105434. https://doi.org/10.1016/j.sysconle.2022.105434 doi: 10.1016/j.sysconle.2022.105434
    [14] W. N. Zhou, J. Yang, X. Q. Yang, A. D. Dai, H. S. Liu, J. A. Fang, Almost surely exponential stability of neural networks with Lévy noise and Markovian switching, Neurocomputing, 145 (2014), 154–159. https://doi.org/10.1016/j.neucom.2014.05.048 doi: 10.1016/j.neucom.2014.05.048
    [15] Q. X. Zhu, Stability analysis of stochastic delay differential equations with Lévy noise, Syst. Control Lett., 118 (2018), 62–68. https://doi.org/10.1016/j.sysconle.2018.05.015 doi: 10.1016/j.sysconle.2018.05.015
    [16] J. Yang, W. N. Zhou, P. Shi, X. Q. Yang, X. H. Zhou, H. Y. Su, Adaptive synchronization of delayed Markovian switching neural networks with Lévy noise, Neurocomputing, 156 (2015), 231–238. https://doi.org/10.1016/j.neucom.2014.12.056 doi: 10.1016/j.neucom.2014.12.056
    [17] C. Imzegouan, Stability for Markovian switching stochastic neural networks with infinite delay driven by Lévy noise, Int. J. Dyn. Control, 7 (2019), 547–556. https://doi.org/10.1007/s40435-018-0451-x doi: 10.1007/s40435-018-0451-x
    [18] P. L. Yu, F. Q. Deng, P. Cheng, Stability analysis of hybrid stochastic delayed Cohen-Grossberg neural networks with Lévy noise and Markov switching, J. Franklin Inst., 359 (2022), 10831–10848. https://doi.org/10.1016/j.jfranklin.2022.05.025 doi: 10.1016/j.jfranklin.2022.05.025
    [19] H. L. Dong, J. Tang, X. R. Mao, Stabilization of highly nonlinear hybrid stochastic differential delay equations with Lévy noise by delay feedback control, SIAM J. Control Optim., 60 (2022), 3302–3325. https://doi.org/10.1137/22M1480392 doi: 10.1137/22M1480392
    [20] X. Y. Li, X. R. Mao, Stabilisation of highly nonlinear hybrid stochastic differential delay equations by delay feedback control, Automatica, 112 (2020), 108657. https://doi.org/10.1016/j.automatica.2019.108657 doi: 10.1016/j.automatica.2019.108657
    [21] Y. Y. Li, R. Dong, X. R. Mao, Discrete-time feedback control for highly nonlinear hybrid stochastic systems with non-differentiable delays, Syst. Control Lett., 175 (2023), 105507. https://doi.org/10.1016/j.sysconle.2023.105507 doi: 10.1016/j.sysconle.2023.105507
    [22] D. Applebaum, Lévy processes and stochastic calculus, 2 Eds., Berlin: Cambridge University Press, 2009. Available from: http://www.cambridge.org/9780521738651.
    [23] C. Zhao, Y. F. Song, Q. X. Zhu, K. B. Shi, Input-to-state stability analysis for stochastic mixed time-delayed neural networks with hybrid impulses, Math. Probl. Eng., 2022, 6135390. https://doi.org/10.1155/2022/6135390
    [24] K. K. Wang, L. Ju, Y. J. Wang, S. H. Li, Impact of colored cross-correlated non-Gaussian and Gaussian noises on stochastic resonance and stochastic stability for a metapopulation system driven by a multiplicative signal, Chaos Soliton. Fract., 108 (2018), 166–181. https://doi.org/10.1016/j.chaos.2018.02.004 doi: 10.1016/j.chaos.2018.02.004
    [25] K. K. Wang, Y. J. Wang, S. H. Li, J. C. Wu, Double time-delays induced stochastic dynamical characteristics for a metapopulation system subjected to the associated noises and a multiplicative periodic signal, Chaos Soliton. Fract., 104 (2017), 400–417. https://doi.org/10.1016/j.chaos.2017.08.030 doi: 10.1016/j.chaos.2017.08.030
  • This article has been cited by:

    1. P. Agilan, K. Julietraja, Nabil Mlaiki, Aiman Mukheimer, Intuitionistic Fuzzy Stability of an Euler–Lagrange Symmetry Additive Functional Equation via Direct and Fixed Point Technique (FPT), 2022, 14, 2073-8994, 2454, 10.3390/sym14112454
    2. Gunaseelan Mani, Arul Joseph Gnanaprakasam, Hüseyin Işık, Fahd Jarad, Fixed point results in C-algebra-valued bipolar metric spaces with an application, 2023, 8, 2473-6988, 7695, 10.3934/math.2023386
    3. Wasfi Shatanawi, Taqi A. M. Shatnawi, New fixed point results in controlled metric type spaces based on new contractive conditions, 2023, 8, 2473-6988, 9314, 10.3934/math.2023468
    4. Jun Fan, Aifeng Tao, Jinhai Zheng, Ji Peng, Numerical Investigation on Temporal Evolution Behavior for Triad Resonant Interaction Induced by Steady Free-Surface Flow over Rippled Bottoms, 2022, 10, 2077-1312, 1372, 10.3390/jmse10101372
    5. Mohammad Abu-Ghuwaleh, Rania Saadeh, Ahmad Qazza, A Novel Approach in Solving Improper Integrals, 2022, 11, 2075-1680, 572, 10.3390/axioms11100572
    6. Rashid Ali, Faisar Mehmood, Aqib Saghir, Hassen Aydi, Saber Mansour, Wajdi Kallel, Solution of integral equations for multivalued maps in fuzzy b-metric spaces using Geraghty type contractions, 2023, 8, 2473-6988, 16633, 10.3934/math.2023851
  • Reader Comments
  • © 2024 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1055) PDF downloads(56) Cited by(0)

Figures and Tables

Figures(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog