Research article Special Issues

Piecewise pseudo almost periodic solutions of interval general BAM neural networks with mixed time-varying delays and impulsive perturbations

  • This paper is concerned with piecewise pseudo almost periodic solutions of a class of interval general BAM neural networks with mixed time-varying delays and impulsive perturbations. By adopting the exponential dichotomy of linear differential equations and the fixed point theory of contraction mapping. The sufficient conditions for the existence of piecewise pseudo almost periodic solutions of the interval general BAM neural networks with mixed time-varying delays and impulsive perturbations are obtained. By adopting differential inequality techniques and mathematical methods of induction, the global exponential stability for the piecewise pseudo almost periodic solutions of the interval general BAM neural networks with mixed time-varying delays and impulsive perturbations is discussed. An example is given to illustrate the effectiveness of the results obtained in the paper.

    Citation: Yanshou Dong, Junfang Zhao, Xu Miao, Ming Kang. Piecewise pseudo almost periodic solutions of interval general BAM neural networks with mixed time-varying delays and impulsive perturbations[J]. AIMS Mathematics, 2023, 8(9): 21828-21855. doi: 10.3934/math.20231113

    Related Papers:

    [1] Jin Gao, Lihua Dai . Weighted pseudo almost periodic solutions of octonion-valued neural networks with mixed time-varying delays and leakage delays. AIMS Mathematics, 2023, 8(6): 14867-14893. doi: 10.3934/math.2023760
    [2] Xiaofang Meng, Yongkun Li . Pseudo almost periodic solutions for quaternion-valued high-order Hopfield neural networks with time-varying delays and leakage delays on time scales. AIMS Mathematics, 2021, 6(9): 10070-10091. doi: 10.3934/math.2021585
    [3] Rakkiet Srisuntorn, Wajaree Weera, Thongchai Botmart . Modified function projective synchronization of master-slave neural networks with mixed interval time-varying delays via intermittent feedback control. AIMS Mathematics, 2022, 7(10): 18632-18661. doi: 10.3934/math.20221025
    [4] Nina Huo, Bing Li, Yongkun Li . Global exponential stability and existence of almost periodic solutions in distribution for Clifford-valued stochastic high-order Hopfield neural networks with time-varying delays. AIMS Mathematics, 2022, 7(3): 3653-3679. doi: 10.3934/math.2022202
    [5] R. Sriraman, P. Vignesh, V. C. Amritha, G. Rachakit, Prasanalakshmi Balaji . Direct quaternion method-based stability criteria for quaternion-valued Takagi-Sugeno fuzzy BAM delayed neural networks using quaternion-valued Wirtinger-based integral inequality. AIMS Mathematics, 2023, 8(5): 10486-10512. doi: 10.3934/math.2023532
    [6] Zhifeng Lu, Fei Wang, Yujuan Tian, Yaping Li . Lag synchronization of complex-valued interval neural networks via distributed delayed impulsive control. AIMS Mathematics, 2023, 8(3): 5502-5521. doi: 10.3934/math.2023277
    [7] Ailing Li, Mengting Lv, Yifang Yan . Asymptotic stability for quaternion-valued BAM neural networks via a contradictory method and two Lyapunov functionals. AIMS Mathematics, 2022, 7(5): 8206-8223. doi: 10.3934/math.2022457
    [8] Huahai Qiu, Li Wan, Zhigang Zhou, Qunjiao Zhang, Qinghua Zhou . Global exponential periodicity of nonlinear neural networks with multiple time-varying delays. AIMS Mathematics, 2023, 8(5): 12472-12485. doi: 10.3934/math.2023626
    [9] Qingyi Cui, Changjin Xu, Wei Ou, Yicheng Pang, Zixin Liu, Jianwei Shen, Muhammad Farman, Shabir Ahmad . Further study on Hopf bifurcation and hybrid control strategy in BAM neural networks concerning time delay. AIMS Mathematics, 2024, 9(5): 13265-13290. doi: 10.3934/math.2024647
    [10] Yongkun Li, Xiaoli Huang, Xiaohui Wang . Weyl almost periodic solutions for quaternion-valued shunting inhibitory cellular neural networks with time-varying delays. AIMS Mathematics, 2022, 7(4): 4861-4886. doi: 10.3934/math.2022271
  • This paper is concerned with piecewise pseudo almost periodic solutions of a class of interval general BAM neural networks with mixed time-varying delays and impulsive perturbations. By adopting the exponential dichotomy of linear differential equations and the fixed point theory of contraction mapping. The sufficient conditions for the existence of piecewise pseudo almost periodic solutions of the interval general BAM neural networks with mixed time-varying delays and impulsive perturbations are obtained. By adopting differential inequality techniques and mathematical methods of induction, the global exponential stability for the piecewise pseudo almost periodic solutions of the interval general BAM neural networks with mixed time-varying delays and impulsive perturbations is discussed. An example is given to illustrate the effectiveness of the results obtained in the paper.



    This paper considers the existence and global exponential stability of piecewise pseudo almost periodic solutions for interval general BAM neural networks with mixed time-varying delays and impulsive perturbations. The mixed time-varying delays include leakage delays and time-varying delays:

    {xi(t)=ai(t)xi(tαi(t))+mj=1sji(t)fj(xj(t),yj(tτji(t)))+ci(t),t>0,ttk,i=1,2,,m,Δxi(t)=xi(t+)xi(t)=Ik(xi(t)),t=tk,kZ+,yj(t)=bj(t)yj(tβj(t))+mi=1tij(t)gi(xi(tδij(t)),yi(t))+dj(t),t>0,ttk,j=1,2,,m,Δyj(t)=yj(t+)yj(t)=Jk(yj(t)),t=tk,kZ+, (1.1)

    where Z+ is the set of nonnegative integers; the numbers of neurons in layers X and Y are denoted by m. xi(t) represents the state variables of the i-th neurons at time t. yj(t) represents the state variables of the j-th neurons at time t. ai(t)>0, bj(t)>0 are continuous functions and represent the decay rates of neurons in different layers, respectively. sji(t), tij(t) are the connection weights. fj(,), gi(,) represent the activation functions of the j-th and i-th units, respectively. ci(t), dj(t) represent the external inputs of the i-th neuron and the j-th neuron acting on different layers at time t, respectively. αi(t)>0, βj(t)>0 are time-varying leakage delays. τji(t)>0, δij(t)>0 are time-varying delays respectively satisfying 1τji(t)>0, 1δij(t)>0. The sequence {tk} has no finite accumulation point. Ik():RmR and Jk():RmR, kZ+.

    The initial value conditions of the system (1.1) are as follows:

    {xi(θ)=φi(θ),θ[r1,0],i=1,2,,m,yj(θ)=ψj(θ),θ[r2,0],j=1,2,,m, (1.2)

    where r1=max1i,jm{suptR+|αi(t)|,suptR+|δij(t)|}, r2=max1i,jm{suptR+|βj(t)|,suptR+|τji(t)|}, φi(θ)C1([r1,0],R), i=1,2,,m, ψj(θ)C1([r2,0],R) and j=1,2,,m.

    BAM neural networks have more complex dynamic behavior. They are composed of two-layer nonlinear feedback networks. Neurons in different layers are interconnected, and there is no connection between neurons in the same layer [1]. It extends the single-layer self-association learning rule to the two-layer hetero-association mode. Time delays in neural networks not only decrease the transmission speed of the networks, but may also change the stability of the networks. There have been many excellent results on the stability and bifurcation behavior of delayed neural networks [2,3,4,5,6,7,8,9]. The existence and stability of solutions for neural networks with time delays have been widely explored [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29]. In particular, the dynamic behavior of neural networks with leakage delays [10,12,13,24,25,27,29] is still a hot research topic.

    Impulses are ubiquitous in the actual neural networks and the dynamic behavior of the neural networks are often affected by impulsive perturbations. And the study of impulsive control theory for neural networks has made a lot of advancements [11,15,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40].

    Compared with the activation functions fj(), gi() in BAM neural networks [10,11,13,14,15,17,19,21,22,23,24], the activation functions fj(xj,yj), gi(xj,yj) in the interval general BAM neural networks [16,17,20,21,22,24] more clearly show the connections between different layers of neurons in the networks. Ding and Huang [16] studied the existence and global robust exponential stability of equilibrium points for the following a class of interval general BAM neural networks with constant delays by using fixed-point theory and constructing suitable Lyapunov functions:

    {xi(t)=aixi(t)+mj=1sjifj(xj(t),yj(tτji))+ci,yj(t)=bjyj(t)+mi=1tijgi(xi(tδij),yi(t))+dj. (1.3)

    By variable transformations, fixed point theory and constructing suitable delay differential inequalities, Xu et al. [17] studied the existence, uniqueness and global exponential stability of the equilibrium point of system (1.3) with proportional delays. Given ai=ai(t), bj=bj(t), sji=sji(t) and tij=tij(t) in the system (1.3), the authors of [20] studied the existence and global exponential stability of periodic solutions for the system (1.3) on time scales by using fixed point theory and constructing suitable Lyapunov functions. Given ai=ai(t), bj=bj(t), sji=sji(t), tij=tij(t), ci=ci(t) and dj=dj(t) in the system (1.3), Duan [22] studied the existence and global exponential stability of pseudo almost periodic solutions for system (1.3) by using exponential dichotomy, fixed point theory and inequality techniques. Given ai=ai(t), bj=bj(t), sji=sji(t), tij=tij(t), ci=ci(t) and dj=dj(t) in the system (1.3), the authors of [24] studied the existence and exponential stability of almost periodic solutions for system (1.3) with leakage delays by using exponential dichotomy, fixed point theory and inequality techniques.

    However, in the existing literature, we have not found any research on the dynamic behavior of interval general BAM neural networks with impulses and leakage delays. The stability of the interval general BAM neural networks may be destroyed by external perturbations and leakage delays. Therefore, the effects of leakage delays and impulsive perturbations on the dynamic behavior of interval general BAM neural networks are worth exploring.

    Motivated by the above discussions, this paper studies the existence and global exponential stability of piecewise pseudo almost periodic solutions for the interval general BAM neural network described by (1.1) with mixed time-varying delays and impulsive perturbations. The mixed time-varying delays include leakage delays and time-varying delays.

    Throughout this paper, let e+i=suptR+|ei(t)|, ei=inftR+|ei(t)|, e+ij=suptR+|eij(t)| and eij=inftR+|eij(t)|. Let T be the set of the real sequences {tk}kZ+ such that σ_=infkZ+t1k=infkZ+(tk+1tk)>0, so limk+tk=+. For {tk}kZ+T, assume that {tjk:tjk=tk+jtk,k,jZ+} are equipotentially almost periodic and t0=0; it can be easily proved that the sequence {tjk} satisfies tjk+itjk=tik+jtik and tjktik=tjik+i. And suppose that the following conditions hold.

    (H1) For u1, v1, u2, v2R, there are nonnegative constants Pj, Qj, Hi and Ki such that |fj(u1,v1)fj(u2,v2)|Pj|u1u2|+Qj|v1v2| and |gi(u1,v1)gi(u2,v2)|Hi|u1u2|+Ki|v1v2|.

    (H2) ai(t), bj(t)C(R+,R+) are almost periodic functions, and suptR+|ai(t)|=a+i, inftR+|ai(t)|=ai, suptR+|bj(t)|=b+j, inftR+|bj(t)|=bj. sji(t), tij(t), ci(t), dj(t):R+R are piecewise pseudo almost periodic functions, and suptR+|sji(t)|=s+ji, suptR+|tij(t)|=t+ij, suptR+|ci(t)|=c+i and suptR+|dj(t)|=d+j.

    (H3) Ik, JkPAP(Z+,Rn) and there are two positive constants I, J satisfying 0<I, J<1 such that |Ik(u1)Ik(u2)|I|u1u2|, |Jk(u1)Jk(u2)|J|u1u2| for u1, u1R, kZ+.

    Remark 1.1. In the existing literature on BAM neural networks, in addition to the condition (H1), the activation functions fj() and gi() are required to satisfy fj(0)=0 and gi(0)=0. In this paper, the activation functions fj(0,0)=0, gi(0,0)=0 are not required, as it only needs to meet the condition (H1).

    In this paper, first, by using the exponential dichotomy of linear differential equations and the fixed point theory for contraction mapping, the existence of piecewise pseudo almost periodic solutions for system (1.1) satisfying the initial value conditions given by (1.2) is studied. And then by using mathematical methods of induction and inequality techniques, the global exponential stability of piecewise pseudo-almost periodic solutions for system (1.1) satisfying the initial value conditions given by (1.2) will be discussed.

    Remark 1.2. Compared with the systems studied in [11,16,17,20,21,22,24,25], the system (1.1) studied in this paper is more general, as some classical Hopfield neural networks and BAM neural networks are special cases of the system (1.1). Duan [22] studied the effects of constant delays in the activation functions on the dynamic behavior of the system, but they did not consider the effects of leakage delays and impulsive perturbations on the dynamic behavior of the system. The mixed time-varying delays studied in this paper include not only the time-varying delays in the activation functions but also the leakage delays; the influences of impulsive perturbations on the dynamic behavior of the system are also considered. This paper not only considers the norm but also (). Therefore, the results obtained in [21,22,24] are special cases of the research results in this paper.

    This section mainly gives the necessary definitions, lemmas and notations to prove the existence and global exponential stability of the piecewise pseudo almost periodic solutions of the system described by (1.1)–(1.2).

    C1(R,Rn) is the set of continuous functions with a continuous derivative, and

    ϕ1=max{max{φ,φ},max{ψ,ψ}},

    where =max1<i<msuptR||.

    BC(R,Rn) is the set of bounded continuous functions, and (BC(R,Rn),) is a Banach space.

    Let PC(R,Rn) be the set of all piecewise continuous functions ϕ:RRn such that ϕ is countinuous at t for t{tk:kZ+} and that ϕ(t+k), ϕ(tk) exist and ϕ(tk)=ϕ(tk) for kZ+.

    Define L(Z+,Rn)={u:Z+Rn,u=supnZ+|u(n)|} and PAP0(Z+,Rn)={uL(Z+,Rn):limn12nnk=nu(k)=0}.

    Definition 2.1. [34] A function fC(R,Rn) is called almost periodic if for ϵ>0, there is L(ϵ)>0 such that every interval of length L(ϵ) contains at least a number δT(f,ϵ) such that |f(t+δ)f(t)|<ϵ for any tR, where T(f,ϵ)={δ|f(t+δ)f(t)|<ϵ,tR}. The set of all almost periodic functions is represented by AP(R,Rn).

    Definition 2.2. [35] A sequence {un} is called almost periodic if for ϵ>0, there is a natural number L(ϵ) such that for kZ, there is at least one number q in [k,k+L(ϵ)] for which the inequality un+qun<ϵ holds for all nN. The set of all almost periodic sequences is represented by AP(Z,Rn).

    Definition 2.3. [36] A function ϕPC(R,Rn) is called piecewise almost periodic if the following conditions are satisfied.

    ● (1) {tjk}, tjk=tk+jtk, k, jZ are equipotentially almost periodic, that is to say, for any ϵ>0, there is a relatively dense set in R of ϵ-almost periods that are common for any of the sequences {tjk}.

    ● (2) For any ϵ>0, there is a constant δ(ϵ)>0 such that, if t and t belong to the same interval of continuity of ϕ and |tt|<δ, then ϕ(t)ϕ(t)<ϵ.

    ● (3) For any ϵ>0, there is a relatively dense set Ω(ϵ) in R such that, if ρΩ(ϵ), then ϕ(t+ρ)ϕ(t)<ϵ for tR, which satisfies |tti|>ϵ, iZ.

    Denote by APT(R,Rn) the set of piecewise almost periodic functions; UPC(R,Rn) is the set of the function fPC(R,Rn) such that f satisfies the condition (2) in Definition 2.3. Let

    PC0T(R,Rn)={ϕPC(R,Rn),limtϕ(t)=0},
    PAP0T(R,Rn)={ϕPC(R,Rn),limr12rrrϕdt=0}.

    Obviously, PC0T(R,Rn)PAP0T(R,Rn).

    Definition 2.4. fPC(R×Ω1×Ω2,Rn) is said to be piecewise almost periodic in t uniformly in (φ,ψ)Ω1×Ω2 if for every compact set K1×K2Ω1×Ω2, {f(,φ,ψ):(φ,ψ)K1×K2} is uniformly bounded, and for ϵ>0, there exists a relatively dense set Ω(ϵ) such that f(t+τ,φ,ψ)f(t,φ,ψ)ϵ for (φ,ψ)K1×K2, τΩ(ϵ), and given tR, |tti|>ϵ.

    Definition 2.5. [37] Sequence {un}nZL(Z,Rn) is called pseudo almost periodic if there exist u1nAP(Z,Rn) and u2nPAP0(Z,Rn) such that un=u1n+u2n. The set of pseudo almost periodic sequences is denoted by PAP(Z,Rn).

    Definition 2.6. [38] A function ϕPC(R,Rn) is known as piecewise pseudo almost periodic if it can be disintegrated into ϕ=ϕ1+ϕ2, where ϕ1APT(R,Rn) and ϕ2PAP0T(R,Rn). The set of piecewise pseudo almost periodic functions is denoted by PAPT(R,Rn).

    Lemma 2.1. [38] Let the sequence of vector-valued functions {Ii}iZ be pseudo almost periodic and there be a constant L>0 such that Ii(u)Ii(v)Luv for all u,vΩ,iZ if ϕPAPT(R,X)UPC(R,X) such that R(ϕ)Ω; then Ii(ϕ(ti)) is pseudo almost periodic.

    Lemma 2.2. [39] If φ()PAPT(R,Rn), then φ(h)PAPT(R,Rn).

    Lemma 2.3. [39] If φ(),ψ()PAPT(R,R), then φ×ψPAPT(R,R).

    Lemma 2.4. [39] If fi()C(R,Rn) satisfies the Lipschitz condition, ψ()PAPT(R,R), ψ()PAPT(R,R) and δ()APT(R,R) such that for all tR, 1δ(t)>0, fi(ψ(δ()))PAPT(R,R).

    Lemma 2.5. [40] Assume that fPAPT(R×Ω1×Ω2,X) and the following conditions hold:

    ● (1) {f(t,u,v):tR,uK1,vK2} is bounded for every bounded subset K1×K2Ω1×Ω2.

    ● (2) {f(t,u,v)} is uniformly continuous in each bounded subset of Ω1×Ω2 uniformly in tR.

    If φPAPT(R,X), ψPAPT(R,X) such that R(φ)×R(ψ)Ω1×Ω2, then f(t,φ(),ψ())PAPT(R,X), where R(φ), R(ψ) are the ranges of φ, ψ, respectively.

    Lemma 2.6. [40] Let fPAPT(R×Ω1×Ω2,X), φPAPT(R,X), ψPAPT(R,X) and R(φ)×R(ψ)Ω1×Ω2. Assume that there exists a constant Lf>0 such that f(t,u1,v1)f(t,u2,v2)Lf(u1u2+v1v2), tR, u1, u2Ω1 and v1, v2Ω2. Then, f(,φ(),ψ())PAPT(R,X).

    Definition 2.7. [34] Let xRn and A() be an n×n continuous matrix defined on R. The linear system x(t)=A(t)x(t) is said to admit an exponential dichotomy on R if there exist positive constants k and α and a projection P such that the fundamental solution matrix X(t) of x(t)=A(t)x(t) satisfies the following:

    X(t)PX1(s)keα(ts),ts,X(t)(IP)X1(s)keα(st),ts.

    Lemma 2.7. [34] Let ai() be an almost periodic function on R and M[ai]=limT+t+Ttai(s)ds>0, i=1,2,...,n. Then, the system x(t)=diag(a1(t),a2(t),...,an(t))x(t) admits an exponential dichotomy on R.

    Lemma 2.8. [41] Let A() be an almost periodic matrix function and f()PAP(R,Rn). If x(t)=A(t)x(t) admits an exponential dichotomy, then the pseudo almost periodic system x(t)=A(t)x(t)+f(t) has a unique pseudo almost periodic solution x(t), and

    x(t)=tX(t)PX1(s)f(s)dstX(t)(IP)X1(s)f(s)ds.

    Definition 2.8. Let ¯z(t)=(¯x(t),¯y(t))T be the solution of system (1.1) satisfying the initial value conditions ¯ϕ=(¯φ(θ),¯ψ(θ))T. If there exist constants M1, η>0 such that for any solution z(t)=(x(t),y(t))T of the system (1.1) satisfying the initial value conditions ϕ=(φ(θ),ψ(θ))T, the following inequality holds

    z(t)¯z(t)1Mϕ¯ϕ1eηt,t0,

    where ϕ¯ϕ1=max{max{φ,φ},max{ψ,ψ}}. Then, ¯z(t) is said to be globally exponentially stable.

    This section establishes the conditions of the existence for piecewise pseudo almost periodic solutions of the system described by (1.1)–(1.2).

    Lemma 3.1. Let fPAPT(R×Ω1×Ω2,X), φ, ψPAPT(R,Rn) and R(φ)×R(ψ)Ω1×Ω2. There are constants P>0, Q>0 such that for tR, φ, ¯φΩ1, ψ, ¯ψΩ2, |f(t,φ,ψ)f(t,¯φ,¯ψ)|P|φ¯φ|+Q|ψ¯ψ| and θ()C(R,R). Then, f(t,φ(tθ()),ψ), f(t,φ,ψ(tθ()))PAPT(R,X) and f(t,φ(tθ()),ψ(tθ()))PAPT(R,X).

    Proof. Since φ, ψPAPT(R,R), then φ=φ1+φ2, ψ=ψ1+ψ2, where φ1, ψ1APT(R,R) and φ2, ψ2PAP0T(R,R). We have

    f(t,φ(tθ()),ψ)=f(t,φ1(tθ())+φ2(tθ()),ψ1+ψ2)=f(t,φ1(tθ()),ψ1)+f(t,φ1(tθ())+φ2(tθ()),ψ1+ψ2)f(t,φ1(tθ()),ψ1)=f1+f2,

    where f1=f(t,φ1(tθ()),ψ1), f2=f(t,φ1(tθ())+φ2(tθ()),ψ1+ψ2)f(t,φ1(tθ()),ψ1). From Lemmas 2.2–2.5, f1=f(t,φ1(tθ()),ψ1)APT(R,R) holds. Now f2PAP0T(R,R) will be proved. Since

    limr12rrr|f2|dtlimr12rrr(P|φ2(tθ(t)|+Q|ψ2|)dt=Plimr12rrr|φ2(tθ(t)|dt+Qlimr12rrr|ψ2|dt=0.

    Thus f2PAP0T(R,R). So, f(t,φ(tθ()),ψ)PAPT(R,X). In a similar way, f(t,φ,ψ(tθ()))PAPT(R,X) and f(t,φ(tθ()),ψ(tθ()))PAPT(R,X). The proof is complete.

    Theorem 3.1. If conditions (H1)(H3) are true. For any ϕ=(φ,ψ)TPAPT(R+,R2m) and ¯ϕ=(¯φ,¯ψ)TPAPT(R+,R2m), define the operator

    Φ1ϕ(t)=(tetsa1(u)du¯F1(s,φ,ψ)ds...tetsam(u)du¯Fm(s,φ,ψ)dstetsb1(u)du¯G1(s,φ,ψ)ds...tetsbm(u)du¯Gm(s,φ,ψ)ds),

    where

    ¯Fi(t,φ,ψ)=ai(t)ttαi(t)φi(s)ds+mj=1sji(t)fj[φj(t),ψj(tτji(t))]+ci(t),i=1,2,,m,¯Gj(t,φ,ψ)=bj(t)ttβj(t)ψj(s)ds+mi=1tij(t)gi[φi(tδij(t)),ψi(t)]+dj(t),j=1,2,,m.

    Then, Φ1ϕ() is a mapping from PAPT(R+,R2m) to PAPT(R+,R2m).

    Proof. By Lemmas 2.2–2.5 and Lemma 3.1, ¯Fi(t,φ,ψ)PAPT(R+,R); thus,

    ¯Fi(t,φ,ψ)=¯F1i(t,φ,ψ)+¯F2i(t,φ,ψ),

    where ¯F1i(t,φ,ψ)APT(R+,R), ¯F2i(t,φ,ψ)PAP0T(R+,R).

    For i=1,2,...,m, we have

    tetsai(u)du¯Fi(s,φ,ψ)ds=tetsai(u)du¯F1i(s,φ,ψ)ds+tetsai(u)du¯F2i(s,φ,ψ)ds=Ψ1(t)+Ψ2(t).

    Step 1: Prove that Ψ1(t)UPC(R+,R). Obviously, Ψ1(t)PC(R+,R). For t,t(tk,tk+1),kZ+,t<t,

    |Ψ1(t)Ψ1(t)|=|tetsai(u)du¯F1i(s,φ,ψ)dstetsai(u)du¯F1i(s,φ,ψ)ds||t(etsai(u)duetsai(u)du)¯F1i(s,φ,ψ)ds|+|ttetsai(u)du¯F1i(s,φ,ψ)ds||ettai(u)du1|tetsai(u)du|¯F1i(s,φ,ψ)|ds+ttetsai(u)du|¯F1i(s,φ,ψ)|dsa+i(tt)teai(ts)|¯F1i(s,φ,ψ)|ds+tteai(ts)|¯F1i(s,φ,ψ)|ds(a+iai(tt)+tteai(ts)ds)¯F1i(s,φ,ψ).

    For ϵ>0, there is 0<δ<min{aiϵ2¯F1ia+i,ϵ2¯F1i} such that for t, t satisfying 0<|tt|<δ, |Ψ1(t)Ψ1(t)|<ϵ holds. So, Ψ1(t)UPC(R+,R).

    Step 2: Prove that Ψ1()APT(R+,R). Since ¯F1i(t,φ,ψ)APT(R+,R), for ϵ>0, there is a relatively dense set Ω(ϵ); for τΩ(ϵ), tR+, |ttk|>ϵ, iZ+ such that |¯F1i(s+τ,φ(s+τ),ψ(s+τ))¯F1i(s,φ(s),ψ(s))|<aiϵ. Then

    |Ψ1(t+τ)Ψ1(t)|=|t+τet+τsai(u)du¯F1i(s,φ,ψ)dstetsai(u)du¯F1i(s,φ,ψ)ds||tet+τs+τai(u)du¯F1i(s+τ,φ(s+τ),ψ(s+τ))dstetsai(u)du¯F1i(s,φ(s),ψ(s))ds||teai(ts)¯F1i(s+τ,φ(s+τ),ψ(s+τ))dsteai(ts)¯F1i(s,φ(s),ψ(s))ds|teai(ts)|¯F1i(s+τ,φ(s+τ),ψ(s+τ))¯F1i(s,φ(s),ψ(s))|ds<ϵ.

    Consequently, Ψ1()APT(R,R).

    Step 3: Prove that Ψ2()PAP0T(R+,R). Since ¯F2i(su,φ(su),ψ(su)),¯F2i(s,φ(s),ψ(s))PAP0T(R+,R), then

    limr12rrr|Ψ2()|dt=limr12rrr|tetsai(u)du¯F2i(s,φ,ψ)ds|dtlimr12rrr|teai(ts)¯F2i(s,φ,ψ)ds|dtlimr12rrreaitdtreais¯F2i(s,φ,ψ)ds+limr12rrrtreai(ts)|¯F2i(s,φ,ψ)|dsdtlimr¯F2i(s,φ,ψ)2rairreai(t+r)dt+limr12rrrt+r0eaiu|¯F2i(tu,φ(tu),ψ(tu))|dudtlimr¯F2i(s,φ,ψ)2rairreai(t+r)dt+0eaiu(limr12rrr|¯F2i(tu,φ(tu),ψ(tu))|dt)du=0.

    Consequently, Ψ2()PAP0T(R+,R). So, tetsai(u)du¯Fi(s,φ,ψ)dsPAPT(R+,R). Similarly, for j=1,2,...,m, we have that tetsbj(u)du¯Gj(s,φ,ψ)dsPAPT(R+,R).

    Therefore Φ1ϕ()PAPT(R+,R2m).

    Theorem 3.2. If conditions (H1)(H3) are true, for any ϕ=(φ,ψ)TPAPT(R+,R2m), define the operator

    Φ2ϕ(t)=(tk<tettka1(u)duIk(φ1(tk))...tk<tettkam(u)duIk(φm(tk))tk<tettkb1(u)duJk(ψ1(tk))...tk<tettkbm(u)duJk(ψm(tk))).

    Then, Φ2ϕ() is a mapping from PAPT(R+,R2m) to PAPT(R+,R2m).

    Proof. Assume that t(tl,tl+1], lZ+; therefore, ttk=ttl+tltk(lk)σ_, and

    tk<teai(ttk)0j=lk<eaijσ_11eaiσ_.

    For any ϵ>0, take the positive constant δ<ϵ(1eaiσ_)a+iIk(φi(tk)), for t,t(tk,tk+1),kZ+,t<t, provided that |tt|<δ; there is ζ(t,t) such that

    |tk<tettkai(u)duIk(φi(tk))tk<tettkai(u)duIk(φi(tk))|a+itk<ζeai(ζtk)Ik(φi(tk))|tt|a+iIk(φi(tk))1eaiσ_|tt|<ϵ.

    That is, tk<tettkai(u)duIk(φi(tk))UPC(R+,R). By Lemma 2.1 and (H3), Ik(φi(tk))PAP(Z+,R). Let Ik(φi(tk))=I1k+I2k, where I1kAP(Z+,R), I2kPAP0(Z+,R); then,

    tk<tettkai(u)duIk(φi(tk))=tk<tettkai(u)duI1k(φi(tk))+tk<tettkai(u)duI2k(φi(tk))=¯I1()+¯I2().

    First, we show that ¯I1()APT(R+,R). Since {tjk:k,jZ+} are equipotentially almost periodic, for any ϵ>0, there is a relatively dense set Ω(ϵ) of R+ and Q(ϵ) of Z+ such that for t(tk,tk+1], |ttk|>ϵ, |ttk+1|>ϵ, kZ+, τΩ(ϵ), qQ(ϵ), from Lemma 35 (see [35]), we have that t+τ>tk+τ+ϵ>tk+q and tk+q+1>tk+1ϵ+τ>t+τ; so, tk+q<t+τ<tk+q+1. Then

    |¯I1(t+τ)¯I1(t)|=|tk<t+τet+τtkai(u)duI1k(φi(tk))tk<tettkai(u)duI1k(φi(tk))|tk<tettkai(u)du|I1k+q(φi(tk))I1k(φi(tk))|ϵtk<teai(ttk)ϵ1eaiσ_.

    So, ¯I1()APT(R+,R).

    Next, we show that ¯I2()PAP0T(R+,R). For k, nZ+ and t(tk,tk+1], let

    χn(t)={ettknai(u)duI2kn(φi(tkn)),0nk,0,n>k;

    since

    limt|ettknai(u)duI2kn(φi(tkn))|limteai(ttkn)I2k(φi(tk))=0,

    limt|χn(t)|=0; there is χn(t)PC0T(R,Rn)PAP0T(R,Rn).

    Let Sn(t)=nl=0χl(t), because

    |sn(t)|nl=0eai(ttkl)I2k(φi(tk))nl=0eai(ttk)elσ_aiI2k(φi(tk)).

    Therefore, limt|sn(t)|=0. So, sn(t)PAP0T(R,Rn).

    And because limnsn(t)=¯I2() uniformly holds on R+, from Lemma 2.2 in [40], ¯I2()PAP0T(R+,R) holds. Therefore, tk<tettkai(u)duIk(φi(tk))PAPT(R+,R).

    In the same way, tk<tettkbj(u)duJk(ψj(tk))PAPT(R+,R) can be obtained. Thus Φ2ϕ()PAPT(R+,R2m).

    From Theorems 3.1 and 3.2, the following theorem is obtained.

    Theorem 3.3. If conditions (H1)(H3) are true, for any ϕ=(φ,ψ)TPAPT(R+,R2m) and ϕ=(φ,ψ)TPAPT(R+,R2m), define the operator

    Φϕ(t)=Φ1ϕ(t)+Φ2ϕ(t).

    Then, Φϕ(t) maps PAPT(R+,R2m) into PAPT(R+,R2m), and Φϕ(t) maps PAPT(R+,R2m) into PAPT(R+,R2m).

    Theorem 3.4. Assume that conditions (H1)(H3) are true and the following condition holds:

    (H4) pij=max{max{θiai+I1eaiσ_,θiai(a+i+ai)+Ia+i1eaiσ_},max{γjbj+J1ebjσ_,γibj(b+j+bj)+Ib+j1ebjσ_}}<1,

    where θi=a+iα+i+mj=1s+ji(Pj+Qj), γj=b+jβ+j+mi=1t+ij(Hi+Ki), i,j=1,2,,m.

    Then, there is a unique piecewise differentiable pseudo almost periodic solution of system (1.1) in the region

    X0={ϕ=(φ1,φ2,,φm,ψ1,ψ2,,ψm)TϕPAPT(R+,R2m),ϕϕ01pL1p},

    where ϕ0=(φ01,φ02,,φ0m,ψ01,ψ02,,ψ0m)T, and

    L=max{max1im{liai+li,liai(a+i+ai)+a+ili},max1im{ljbj+lj,ljbj(b+j+bj)+b+jlj}},φ0i(t)=tetsai(u)du(ci(s)+mj=1sji(s)fj(0,0))ds+tk<tettkai(u)duIk(0),ψ0j(t)=tetsbj(u)du(dj(s)+mi=1tij(s)gi(0,0))ds+tk<tettkbj(u)duJk(0).

    Proof. It is easy to show that X0PAPT(R+,R2m) is a closed convex set. Rewrite (1.1) into the following form

    {xi(t)=ai(t)xi(t)+ai(t)ttαi(t)xi(s)ds+mj=1sji(t)fj(xj(t),yj(tτji(t)))+ci(t),ttk,i=1,2,,m,Δxi(t)=xi(t+k)xi(tk)=Ik(xi(tk)),t=tk,yj(t)=bj(t)yj(t)+bj(t)ttβj(t)yj(s)ds+mi=1tij(t)gi(xi(tδij(t)),yi(t))+dj(t),ttk,j=1,2,,m,Δyj(t)=yj(t+k)yj(tk)=Jk(yj(tk)),t=tk. (3.1)

    We consider the following system

    {xi(t)=ai(t)xi(t)+¯Fi(t,φ,ψ)+Ik(φi),i=1,2,,m,yj(t)=bj(t)yj(t)+¯Gj(t,φ,ψ)+Jk(ψj),j=1,2,,m, (3.2)

    where

    ¯Fi(t,φ,ψ)=ai(t)ttαi(t)φi(s)ds+mj=1sji(t)fj(φj(t),ψj(tτji(t)))+ci(t),i=1,2,,m,¯Gj(t,φ,ψ)=bj(t)ttβj(t)ψj(s)ds+mi=1tij(t)gi(φi(tδij(t)),ψi(t))+dj(t),j=1,2,,m.

    Since M[ai]>0, M[bi]>0, from Lemma 2.7, we have that the linear system

    {xi(t)=ai(t)xi(t),i=1,2,,m,yj(t)=bj(t)yj(t),j=1,2,,m (3.3)

    admits an exponential dichotomy on R+. Thus, by Lemma 2.8, we obtain that system (3.2) has a unique piecewise pseudo almost periodic solution, which is expressed as follows:

    Φϕ(t)=(tetsa1(u)du¯F1(s,φ,ψ)ds+tk<tettka1(u)duIk(φ1(tk))...tetsam(u)du¯Fm(s,φ,ψ)ds+tk<tettkam(u)duIk(φm(tk))tetsb1(u)du¯G1(s,φ,ψ)ds+tk<tettkb1(u)duJk(ψ1(tk))...tetsbm(u)du¯Gm(s,φ,ψ)ds+tk<tettkbm(u)duJk(ψm(tk))). (3.4)

    Let li=c+i+mj=1s+ji|fj(0,0)|, li=supkZ+|Ik(0)|1eaiσ_, lj=d+j+mi=1t+ij|gi(0,0)| and lj=supkZ+|Jk(0)|1ebjσ_. Then,

    |φ0i(t)|=|tetsai(u)du(ci(s)+mj=1sji(s)fj(0,0))ds+tk<tettkai(u)duIk(0)|teai(ts)(c+i+mj=1s+ji|fj(0,0)|)ds+tk<teai(ttk)|Ik(0)|liai+li.

    Similarly, ψ0j(t)ljbj+lj.

    In addition,

    |(φ0i)(t)|=|ci(t)+mj=1sji(t)fj[0,0]ai(t)tetsai(u)du(ci(s)+mj=1sji(s)fj[0,0])dstk<tai(t)ettkai(u)duIk(0)|li(1+a+iteai(ts)ds)+a+itk<teai(ttk)Ik(0)liai(a+i+ai)+a+ili.

    Similarly, |(ψ0j)(t)|ljbj(b+j+bj)+b+jlj.

    Therefore, ϕ01L. For ϕX0, ϕ1ϕϕ01+ϕ01L1p.

    We will prove that Φϕ:X0X0 is a contraction.

    First, we prove that for any ϕX0 and ΦϕX0, let

    Fi(t,φ,ψ)=ai(t)ttαi(t)φi(s)ds+mj=1sji(t)(fj(φj(t),ψj(tτji(t)))fj(0,0)),Gj(t,φ,ψ)=bj(t)ttβj(t)ψj(s)ds+mi=1tij(t)(gi(φi(tδij(t)),ψi(t))gi(0,0)).

    Then,

    |Fi(t,φ,ψ)|a+iφα+i+mj=1s+ji(|fj(φj(t),ψj(tτji(t)))fj(0,0)|)a+iφα+i+mj=1s+ji(Pj|φj(t)|+Qj|ψj(tτji(t))|)a+iφα+i+mj=1s+ji(Pjφ+Qjψ)(a+iα+i+mj=1s+ji(Pj+Qj))ϕ1=θiϕ1,i=1,2,,m.

    Similarly, |Gj(t,φ,ψ)|(b+jβ+j+mi=1t+ij(Hi+Ki))ϕ1=γjϕ1,j=1,2,,m, where θi=a+iα+i+mj=1s+ji(Pj+Qj),γj=b+jβ+j+mi=1t+ij(Hi+Ki),i,j=1,2,,m.

    For i=1,2,,m, we have

    |(Φϕϕ0)i(t)|=|tetsai(u)duFi(s,φ,ψ)ds+tk<tettkai(u)du(Ik(φi(tk))Ik(0))|θiϕ1te(ts)aids+tk<te(ttk)aiIϕ1=(θiai+I1eaiσ_)ϕ1;

    similarly, for j=1,2,,m, |(Φϕϕ0)j(t)|(γjbj+J1ebjσ_)ϕ1.

    Furthermore, for i=1,2,,m, we have

    |(Φϕϕ0)i(t)|=|(tetsai(u)duFi(s,φ,ψ)ds+tk<tettkai(u)du(Ik(φi(tk))Ik(0)))|=|Fi(t,φ,ψ)ai(t)tetsai(u)duFi(s,φ,ψ)dstk<tai(t)ettkai(u)du(Ik(φi(tk))Ik(0))|(θiai(a+i+ai)+a+iI1eaiσ_)ϕ1.

    Similarly, for j=1,2,,m, we have

    |(Φϕϕ0)j(t)|(γjbj(b+j+bj)+b+jJ1ebjσ_)ϕ1.

    In the light of (H4), we have that Φϕϕ01pϕ1<pL1p, that is ΦϕX0.

    Next, we prove that Φϕ is a contraction. For ϕ=(φ1,φ2,,φm,ψ1,ψ2,,ψm)T,¯ϕ=(¯φ1,¯φ2,,¯φm,¯ψ1,¯ψ2,,¯ψm)TX0, it follows that

    |¯Fi(t,φ,ψ)¯Fi(t,¯φ,¯ψ)|ai(t)ttαi(t)|φi(s)¯φi(s)|ds+mj=1sji(t)|fj[φj(t),ψj(tτji(t))]fj[¯φj(t),¯ψj(tτji(t))]|a+i|φ¯φ|α+i+mj=1s+ji(Pj|φ¯φ|+Qj|ψ¯ψ|)(a+iα+i+mj=1s+ji(Pj+Qj))ϕ¯ϕ1=θiϕ¯ϕ1,i=1,2,,m,

    and |¯Gj(t,φ,ψ)¯Gj(t,¯φ,¯ψ)|(b+jβ+j+mi=1t+ij(Pi+Qi))ϕ¯ϕ1=γjϕ¯ϕ1,j=1,2,,m.

    And because

    |tk<tettkai(u)duIk(φi(tk))tk<tettkai(u)duIk(¯φi(tk))|tk<tettkai(u)du|Ik(φi(tk))Ik(¯φi(tk))|I1eaiσ_ϕ¯ϕ1,

    and

    |(tk<tettkai(u)duIk(φi(tk))tk<tettkai(u)duIk(¯φi(tk)))|ai(t)tk<tettkai(u)du|Ik(¯φi(tk))Ik(φi(tk))|a+iI1eaiσ_ϕ¯ϕ1.

    Similarly,

    |tk<tettkbj(u)duJk(ψj(tk))tk<tettkbj(u)duJk(¯ψj(tk))|I1ebjσ_ϕ¯ϕ1|(tk<tettkbj(u)duJk(ψj(tk))tk<tettkbj(u)duJk(¯ψj(tk)))|b+jI1ebjσ_ϕ¯ϕ1.

    So, for i=1,2,,m, we have

    |ΦϕΦ¯ϕ|itetsai(u)du|¯Fi(s,φ,ψ)¯Fi(s,¯φ,¯ψ)|ds+tk<tettkai(u)du|Ik(φi(tk))Ik(¯φi(tk))|θiaiϕ¯ϕ1+I1eaiσ_ϕ¯ϕ1=(θiai+I1eaiσ_)ϕ¯ϕ1,

    and

    |(ΦϕΦ¯ϕ)|i=(tetsai(u)du(¯Fi(s,φ,ψ)¯Fi(s,¯φ,¯ψ))ds+tk<tettkai(u)du(Ik(φi(tk))Ik(¯φi(tk))))|Fi(t,φ,ψ)Fi(t,¯φ,¯ψ)|+a+itetsai(u)du|Fi(s,φ,ψ)Fi(s,¯φ,¯ψ)|ds+a+iI1eaiσ_ϕ¯ϕ1(θiai(a+i+ai)+a+iI1eaiσ_)ϕ¯ϕ1.

    Similarly, for j=1,2,,m, we have

    |ΦϕΦ¯ϕ|j(γjbj+J1ebjσ_)ϕ¯ϕ1,|(ΦϕΦ¯ϕ)|j(γjbj(b+j+bj)+b+jJ1ebjσ_)ϕ¯ϕ1.

    By (H4), we have

    ΦϕΦ¯ϕ1pijϕ¯ϕ1,

    which implies that Φϕ is a contraction. Therefore, Φϕ has a fixed point in X0, that is, (1.1) has a unique piecewise pseudo almost periodic solution in X0. This completes the proof.

    This section states and proves the sufficient conditions for the global exponential stability of piecewise pseudo almost periodic solutions of the system described by (1.1)–(1.2).

    Theorem 4.1. Assume that conditions (H1)(H4) are true and the following condition holds.

    (H5) For i, j=1,2,,m,

    ai>max{a+iα+i+mj=1s+ji(Pj+Qj),(a+i+ai)[a+iα+i+mj=1s+ji(Pj+Qj)]},bj>max{b+jβ+j+mi=1t+ij(Hi+Ki),(b+j+bj)[b+jβ+j+mi=1t+ij(Hi+Ki)]}.

    Then, the piecewise pseudo almost periodic solutions of the system described by (1.1)–(1.2) is globally exponentially stable.

    Proof. According to Theorem 3.4, the system (1.1) has at least one piecewise differentiable pseudo-almost periodic solution (¯x1(t),¯x2(t),,¯xm(t),¯y1(t),¯y2(t),,¯ym(t))TX0 with the initial value condition ¯ϕ=(¯φ1(θ),¯φ2(θ),,¯φm(θ),¯ψ1(θ),¯ψ2(θ),,¯ψm(θ))T. Let

    (x1(t),x2(t),,xm(t),y1(t),y2(t),,ym(t))T

    be an arbitrary solution of system (1.1) with the initial value condition

    ϕ=(φ1(θ),φ2(θ),,φm(θ),ψ1(θ),ψ2(θ),,ψm(θ))T.

    Let ui()=xi()¯xi(), vj()=yj()¯yj(); for ttk and t>0, we have

    ui(t)=ai(t)ui(t)+ai(t)ttαi(t)ui(s)ds+mj=1sji(t)[fj(xj(t),yj(tτji(t)))fj(¯xj(t),¯yj(tτji(t)))], (4.1)
    vj(t)=bj(t)vj(t)+bj(t)ttβj(t)vj(s)ds+mi=1tij(t)[gi(xi(tδij(t)),yi(t)gi(xi(tδij(t)),yi(t)]. (4.2)

    The initial value conditions are as follows:

    {ui(θ)=φi(θ)¯φi(θ),θ[r1,0],i=1,2,,m,vj(θ)=ψj(θ)¯ψj(θ),θ[r2,0],j=1,2,,m.

    Multiplying estkai(τ)dτ on both sides of (4.1) and estkbj(τ)dτ on both sides of (4.2), and integrating over [tk,t], we derive

    ui(t)=ui(tk)ettkai(τ)dτ+ttketsai(τ)dτ{ai(s)ssαi(s)ui(τ)dτ+mj=1sji(s)[fj(xj(s),yj(sτji(s)))fj(¯xj(s),¯yj(sτji(s)))]}ds, (4.3)
    vj(t)=vj(tk)ettkbj(τ)dτ+ttketsbj(τ)dτ{bj(s)ssβj(s)vj(τ)dτ+mi=1tij(s)[fi(xi(sδij(s)),yi(s))fj(¯xj(sδij(s)),¯yj(s))]}ds. (4.4)

    Consider the following functions:

    Ci(s)=aishi(s),s(0,ai),s(0,ai),Di(s)=ais(a+i+ai)hi(s)+shi(s),s(0,ai),¯Cj(s)=bjs¯hj(s),s(0,bj),¯Dj(s)=bjs(b+j+bj)¯hj(s)+s¯hj(s),s(0,bj),

    where hi(s)=a+iesα+i1s+mj=1s+ji(Pj+Qjesτ+ji), ¯hj(s)=b+jesβ+j1s+mi=1t+ij(Hiesδ+ij+Ki).

    For i,j=1,2,,m, since Ci(s), Di(s) are continuous on (0,ai), ¯Cj(s), ¯Dj(s) are continuous on (0,bj), lims0Ci(s)>0, lims0¯Cj(s)>0, lims0Di(s)>0, lims0¯Dj(s)>0, limsaiCi(s)<0, limsaiDi(s)<0, limsbj¯Cj(s)<0 and limsbj¯Dj(s)<0. There exist constants si1, si2(0,ai) and sj3, sj4(0,bj) such that Ci(si1)=0, Di(si2)=0, ¯Cj(sj3)=0 and ¯Dj(sj4)=0. There exists a constant η(0,min1i,jm{si1,si2,sj3,sj4}) such that Ci(η)>0, Di(η)>0, ¯Cj(η)>0 and ¯Dj(η)>0. Obviously, there is η(0,min{ai,bj}).

    Take a positive sequence {Mk}k=1 satisfying

    1M1<1N0min{hi(η)aiη,¯hj(η)bjη},
    1Mk<1Nk1min{hi(η)aiη,¯hj(η)bjη},

    where Nk=Mkmax{1+I,1+J}, k1, N0=max{1+I,1+J}.

    Let

    w(t)={wi(t)=ui(t),i=1,2,,m,wm+j(t)=vj(t),j=1,2,,m.

    Let Υ=ϕ¯ϕ1; for i,j=1,2,,m, it is obvious that there exists a positive constant M0>1 such that the following inequalities hold

    wi(θ)1M0Υeηt,θ[r1,0],wm+j(θ)1M0Υeηt,θ[r2,0].

    For t=t+0, we have

    |ui(t+0)|=|xi(t+0)¯xi(t+0)||xi(t0)¯xi(t0)|+|I1(xi(t0))I1(¯xi(t0))|=|ui(t0)|+I|ui(t0)|(1+I)Υ,|vj(t+0)|=|yj(t+0)¯yj(t+0)||yj(t0)yj(t0)|+|J1(yj(t0))J1(yj(t0))|=|vj(t0)|+J|vj(t0)|(1+J)Υ.

    Therefore,

    wl(t+0)N0Υ. (4.5)

    Now we will prove, for t(0,t1], l=1,2,,2m, that

    wl(t)1M1Υeηt. (4.6)

    In order to prove that (4.6) is true, we first prove that for any ρ>1, t(0,t1], the following inequality holds:

    wl(t)1<M1ρΥeηt. (4.7)

    If (4.7) is not established, there are l0{1,2,,2m} and t(0,t1] such that

    wl0(t)1M1ρΥeηt,wl(t)1<M1ρΥeηt,t(0,t), (4.8)

    where t(0,t1] is the smallest variable such that (4.8) holds. There is a constant σ1 such that

    wl0(t)1=σM1ρΥeηt (4.9)

    and wl(t)1<M1ρΥeηt,t(0,t).

    For l0{1,2,,m}, let i0=l0; from (H5) and (4.3), we derive

    |ui0(t)|N0Υeai0t+t0etsai0(τ)dτ{a+i0ssαi0(s)M1ρΥeητdτ+mj=1s+ji0(Pj|uj(s)|+Qj|vj(sτji0(s))|)}dsN0Υeai0t+t0etsai0(τ)dτ{a+i0ssαi(s)M1ρΥeητdτ+mj=1s+ji0(PjM1ρΥeηs+QjM1ρΥeη(sτ+ji0))}dsN0Υeai0t+eai0tM1ρΥe(ai0η)t1ai0ηhi0(η)=N0Υeai0tM1ρΥeai0tai0ηhi0(η)+M1ρΥeηtai0ηhi0(η){N0M11ai0ηhi0(η)}eai0tM1ρΥ+1ai0ηhi0(η)M1ρΥeηt<M1ρΥeηt, (4.10)
    |ui0(t)|=|ui0(0)ai0(t)et0ai0(τ)dτai0(t)t0etsai0(τ)dτ{ai0(s)ssαi0(s)ui0(τ)dτ+mj=1sji0(s)[fj(xj(s),yj(sτji0(s)))fj(¯xj(s),¯yj(sτji0(s)))]}ds+et0ai0(τ)dτet0ai0(τ)dτ{ai0(t)ttαi0(t)ui0(τ)dτ+mj=1sji0(t)[fj(xj(t),yj(tτji0(t)))fj(¯xj(t),¯yj(tτji0(t)))]}||ui0(0)|a+i0eai0t+a+i0eai0tt0eai0s{a+i0ssαi0(s)|ui0(τ)|dτ+mj=1s+ji0(Pj|uj(s)|+Qj|vj(sτji0(s))|)}ds+a+ittαi0(t)|ui0(τ)|dτ+mj=1s+ji0(Pj|uj(t)|+Qj|vj(tτji0(t))|)N0Υa+i0eai0t+a+i0eai0tt0eai0s{a+i0ssαi0(s)M1ρΥeητdτ+mj=1s+ji0(PjM1ρΥeηs+QjM1ρΥeηseητ+ji0)}ds+a+i0ttαi0(t)M1ρΥeητdτ+mj=1s+ji0(PjM1ρΥeηt+QjM1ρΥeηteητ+ji0){N0M1a+i0eai0t+a+i0eηteai0tai0ηhi0(η)}M1ρΥ+hi0(η)M1ρΥeηt={N0M1hi0(η)ai0η}a+i0eai0tM1ρΥ+{a+i0ai0ηhi0(η)+hi0(η)}M1ρΥeηt<M1ρΥeηt, (4.11)

    where hi0(η)=a+i0eηα+i01η+mj=1s+ji0(Pj+Qjeητ+ji0). Equations (4.10) and (4.11) are both in conflict with (4.9).

    For l0{m+1,m+2,,2m}, let j0=l0m; similarly, from (H5) and (4.4), we derive

    |vj0(t)|{N0M11bj0η¯hj0(η)}ebj0tM1ρΥ+1bj0η¯hj0(η)M1ρΥeηt<M1ρΥeηt, (4.12)
    |vj0(t)|={N0M1¯hj0(η)bj0η}b+j0ebj0tM1ρΥ+{b+j0bj0η¯hj0(η)+¯hj0(η)}M1ρΥeηt<M1ρΥeηt, (4.13)

    where ¯hj0(η)=b+j0eηβ+j01η+mi=1t+ij0(Hieηδ+ij0+Ki). Equations (4.12) and (4.13) are both in conflict with (4.9).

    Therefore, (4.8) is not valid. For t(0,t1], we have that (4.7) holds. Let ρ1; for t(0,t1], (4.6) holds.

    For t=t+1, we have

    |ui(t+1)|=|xi(t+1)¯xi(t+1)||xi(t1)¯xi(t1)|+|I1(xi(t1))I1(¯xi(t1))|=|ui(t1)|+I|ui(t1)|(1+I)M1Υeηt1,|vj(t+1)|=|yj(t+1)¯yj(t+1)||yj(t1)yj(t1)|+|J1(yj(t1))J1(yj(t1))|=|vj(t1)|+J|vj(t1)|(1+J)M1Υeηt1.

    Therefore,

    wl(t+1)N1Υeηt1. (4.14)

    For t(t1,t2], l=1,2,,2m, we certify that

    wl(t)1M2Υeηt. (4.15)

    In order to prove that (4.15) is true, we first prove that for any ρ>1 and t(t1,t2], the following inequality holds:

    wl(t)1<M2ρΥeηt. (4.16)

    If (4.16) is not established, there are l1{1,2,,2m} and t(t1,t2] such that

    wl1(t)1M2ρΥeηt,wl(t)1<M2ρΥeηt,t(t1,t), (4.17)

    where t(t1,t2] is the smallest variable such that (4.17) holds. There is a constant σ11 such that

    wl1(t)1=σ1M2ρΥeηt (4.18)

    and wl(t)1<M2ρΥeηt,t(t1,t).

    For l1{1,2,,m}, let i1=l1; from (H5) and (4.3), we derive

    |ui1(t)|N1ρΥeηt1eai1(tt+1)+tt+1etsai1(τ)dτ{a+i1ssαi1(s)M2ρΥeητdτ+mj=1s+ji1(Pj|uj(s)|+Qj|vj(sτji1(s))|)}dsN1ρΥeηt1eai1(tt+1)+eai1tM2ρΥe(ai1η)te(ai1η)t1ai1ηhi1(η)=N1ρΥeηt1eai1(tt+1)+M2ρΥeηteai1(tt1)eηt1ai1ηhi1(η){N1M2hi1(η)ai1η}eai1(tt1)M2ρΥeηt1+1ai1ηhi1(η)M2ρΥeηt<M2ρΥeηt, (4.19)
    |ui1(t)||ui1(t+1)|a+i1eai1(tt+1)+a+i1eai1ttt+1eai1s{a+i1ssαi1(s)|ui1(τ)|dτ+mj=1s+ji1(Pj|uj(s)|+Qj|vj(sτji1(s))|)}ds+a+ittαi1(t)|ui1(τ)|dτ+mj=1s+ji1(Pj|uj(t)|+Qj|vj(tτji1(t))|)N1ρΥeηt1a+i1eai1(tt1)+a+i1eai1ttt+1eai1s{a+i1ssαi1(s)M2ρΥeητdτ+mj=1s+ji1(PjM2ρΥeηs+QjM2ρΥeηseητ+ji1)}ds+a+i1ttαi1(t)M2ρΥeητdτ+mj=1s+ji1(PjM2ρΥeηt+QjM2ρΥeηteητ+ji1){N1M2hi1(η)ai1η}a+i1eai1(tt1)M2ρΥeηt1+{a+i1ai1ηhi1(η)+hi1(η)}M2ρΥeηt<M2ρΥeηt, (4.20)

    where hi1(η)=a+i1eηα+i11η+mj=1s+ji1(Pj+Qjeητ+ji1). Equations (4.19) and (4.20) are both in conflict with (4.18).

    For l1{m+1,m+2,,2m}, let j1=l1m; similarly, from (H5) and (4.4), we derive

    |vj1(t)|{N1M2¯hj1(η)bj1η}ebj1(tt1)M2ρΥeηt1+1bj1η¯hj1(η)M2ρΥeηt<M2ρΥeηt, (4.21)
    |vj1(t)|{N1M2¯hj1(η)bj1η}b+j1ebj1(tt1)M2ρΥeηt1+{b+j1bj1η¯hj1(η)+¯hj1(η)}M2ρΥeηt<M2ρΥeηt, (4.22)

    where ¯hj1(η)=b+j1eηβ+j11η+mi=1t+ij1(Hieηδ+ij1+Ki). Equations (4.21) and (4.22) are both in conflict with (4.18).

    Therefore, (4.17) is not valid. For t(t1,t2], we have that (4.16) holds. Let ρ1; for t(t1,t2], (4.15) holds.

    Similar to the proof of (4.14) and (4.15), we deduce that

    wl(t+k1)Nk1Υeηtk1,wl(t)1MkΥeηt,t(tk1,tk]. (4.23)

    Let M=supk1max{Mk,Nk1}; obviously, M>1; we have

    wl(t)1MΥeηt,t>0. (4.24)

    Therefore, for i,j=1,2,,m,

    xi(t)¯xi(t)1MΥeηt,t>0,yj(t)¯yj(t)1MΥeηt,t>0. (4.25)

    From Definition 2.8, the piecewise pseudo almost periodic solution

    (¯x1(t),¯x2(t),,¯xm(t),¯y1(t),¯y2(t),,¯ym(t))T

    of system (1.1) with the initial value conditions

    ¯ϕ=(¯φ1(θ),¯φ2(θ),,¯φm(θ),¯ψ1(θ),¯ψ2(θ),,¯ψm(θ))T

    is said to be globally exponentially stable.

    This section presents an example to illustrate the effectiveness of our results obtained in previous sections. Consider the following interval general BAM neural network with mixed time-varying delays and impulsive perturbations:

    {xi(t)=ai(t)xi(tαi(t))+2j=1sji(t)fj(xj(t),yj(tτji(t)))+ci(t),t>0,t2k,i=1,2,Δxi(2k)=Ik(xi(2k)),kZ+,yj(t)=bj(t)yj(tβj(t))+2i=1tij(t)gi(xi(tδij(t)),yi(t))+dj(t),t>0,t2k,j=1,2,Δyj(2k)=Jk(yj(2k)),kZ+, (5.1)

    where fj(xj(t),yj(tτji(t)))=|xj(t)|+|yj(tτji(t)|,gi(xi(tδij(t)),yi(t))=|xi(tδij(t))|+|yi(t)|,i,j=1,2, and

    a(t)=(0.2+0.1cos2t0.2+0.1sin4t),b(t)=(0.2+0.1cos2t0.2+0.1sin4t),s(t)=(180+180sint180+180cost180+180sint180+180cost);
    t(t)=(180+180sint180+180cost180+180sint180+180cost),c(t)=(0.5+0.1sint10(1+t2)0.2+0.1cost10(1+t2)),d(t)=(0.3+0.1sint10(1+t2)0.4+0.1cost10(1+t2));
    α(t)=(18|sint|18|cost|),β(t)=(18|cost|18|sint|),δ(t)=(0.4|sint|0.2|cost|0.3|cost|0.1|sint|);
    τ(t)=(0.5|cost|0.3|sint|0.2|sint|0.4|cost|),(Δx1(2k)Δx2(2k)Δy1(2k)Δy2(2k))=(180x1(2k)+180sin(x1(2k))+4180x2(2k)+180cos(x2(2k))+3180y1(2k)+180sin(y1(2k))+2180y2(2k)+180cos(y2(2k))+1.5).

    Obviously,

    a+=(0.30.3),a=(0.20.2),b+=(0.30.3),b=(0.20.2),α+=(1818),s+=(280280280280);
    β+=(1818),t+=(280280280280),δ+=(0.40.20.30.1),τ+=(0.50.30.20.4).

    By calculation, it is easy to acquire σ_=2, Pj=Qj=Hi=Ki=1, i,j=1,2,,m, I=J=140 and

    (θ1θ2γ1γ2)=(1180118011801180),(I1ea1σ_I1ea2σ_J1eb1σ_J1eb2σ_)=(0.07580.07580.07580.0758).

    For i,j=1,2,,m, pij=0.7633<1 and 0.2>1180. That is, the conditions (H1)(H5) are valid. From Theorems 3.4 and 4.1, system (5.1) has a piecewise pseudo almost periodic solution, which is globally exponentially stable. Computer simulations of the existence and stable behavior of the pseudo-almost periodic solutions are given in Figures 1 and 2, respectively.

    Figure 1.  The behavior of x, y for system (5.1) satisfying an initial value condition.
    Figure 2.  The behavior of x, y for system (5.1) satisfying the random 10 initial value conditions, respectively.

    The pseudo almost periodic behaviors have been applied to the qualitative theory of differential equations. Piecewise pseudo almost periodic solutions of BAM neural networks are generalizations of almost periodic solutions, which have obvious scientific significance and application value in many fields such as signal processing, pattern recognition, associative memory, image processing and optimization problems.

    This paper investigated the piecewise pseudo almost periodic solutions of the interval general BAM neural networks with mixed time-varying delays and impulsive perturbations. By employing the exponential dichotomy of linear differential equations, fixed-point theory for contraction mapping, differential inequality techniques and mathematical methods of induction, the effective conditions for the existence and global exponential stability of piecewise pseudo almost periodic solutions of system (1.1) have been established. From Theorems 3.4 and 4.1, the delays contained in the activation functions do not affect the existence and global exponential stability of piecewise pseudo almost periodic solutions of system (1.1). The existence and global exponential stability of the piecewise pseudo almost periodic solutions of the system (1.1) are determined by the negative feedback terms, leakage delays, connection weights, impulsive perturbations and the Lipschitz constants of the activation functions.

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

    The authors declare that they have no competing interests concerning the publication of this article.



    [1] B. Kosko, Bidirectional associative memories, In: IEEE Transactions on Systems, Man and Cybernetics, New York: IEEE, 1988, 49–60. http://doi.org/10.1109/21.87054
    [2] C. D. Huang, J. Wang, X. P. Chen, J. D. Cao, Bifurcations in a fractional-order BAM neural network with four different delays, Neural Networks, 141 (2021), 344–354. http://doi.org/10.1016/j.neunet.2021.04.005 doi: 10.1016/j.neunet.2021.04.005
    [3] C. J. Xu, M. X. Liao, P. L. Li, Y. Guo, Z. X. Liu, Bifurcation properties for fractional order delayed BAM neural networks, Cogn. Comput., 13 (2021), 322–356. http://doi.org/10.1007/s12559-020-09782-w doi: 10.1007/s12559-020-09782-w
    [4] C. J. Xu, Z. X. Liu, P. L. Li, J. L. Yan, L. Y. Yao, Bifurcation mechanism for fractional-order three-triangle multi-delayed neural network, Neural Process. Lett., 2022 (2022), 1–27. http://doi.org/10.1007/s11063-022-11130-y doi: 10.1007/s11063-022-11130-y
    [5] C. J. Xu, D. Mu, Z. X. Liu, Y. C. Pang, M. X. Liao, P. L. Li, et al., Comparative exploration on bifurcation behavior for integer-order and fractional-order delayed BAM neural networks, Nonlinear Anal. Model., 27 (2022), 1030–1053. http://doi.org/10.15388/namc.2022.27.28491 doi: 10.15388/namc.2022.27.28491
    [6] H. S. Hou, H. Zhang, Stability and hopf bifurcation of fractional complex-valued BAM neural networks with multiple time delays, Appl. Math. Comput., 450 (2023), 127986. http://doi.org/10.1016/j.amc.2023.127986 doi: 10.1016/j.amc.2023.127986
    [7] C. J. Xu, D. Mu, Z. X. Liu, Y. C. Pang, M. X. Liao, C. Aouiti, New insight into bifurcation of fractional-order 4D neural networks incorporating two different time delays, Commun. Nonlinear Sci., 118 (2023), 107043. http://doi.org/10.1016/j.cnsns.2022.107043 doi: 10.1016/j.cnsns.2022.107043
    [8] Q. R. Dai, Exploration of bifurcation and stability in a class of fractional-order super-double-ring neural network with two shared neurons and multiple delays, Chaos Soliton. Fract., 168 (2023), 113185. http://doi.org/10.1016/j.chaos.2023.113185 doi: 10.1016/j.chaos.2023.113185
    [9] C. J. Xu, W. Zhang, C. Aouiti, Z. X. Liu, L. Y. Yao, Bifurcation insight for a fractional-order stage-structured predator-prey system incorporating mixed time delays, Math. Method. Appl. Sci., 46 (2023), 9103–9118. https://doi.org/10.1002/mma.9041 doi: 10.1002/mma.9041
    [10] Y. K. Li, Y. Q. Li, Existence and exponential stability of almost periodic solution for neutral delay BAM neural networks with time-varying delays in leakage terms, J. Franklin I., 350 (2013), 2808–2825. http://doi.org/10.1016/j.jfranklin.2013.07.005 doi: 10.1016/j.jfranklin.2013.07.005
    [11] C. Wang, Almost periodic solutions of impulsive BAM neural networks with variable delays on time scales, Commun. Nonlinear Sci., 19 (2014), 2828–2842. https://doi.org/10.1016/j.cnsns.2013.12.038 doi: 10.1016/j.cnsns.2013.12.038
    [12] Y. K. Li, L. Yang, B. Li, Existence and stability of pseudo almost periodic solution for neutral type high-order hopfield neural networks with delays in leakage terms on time scales, Neural Process. Lett., 44 (2016), 603–623. https://doi.org/10.1007/s11063-015-9483-9 doi: 10.1007/s11063-015-9483-9
    [13] C. Aouiti, I. B. Gharbia, J. D. Cao, M. S. M'hamdi, A. Alsaedi, Existence and global exponential stability of pseudo almost periodic solution for neutral delay BAM neural networks with time-varying delay in leakage terms, Chaos Soliton. Fract., 107 (2018), 111–127. https://doi.org/10.1016/j.chaos.2017.12.022 doi: 10.1016/j.chaos.2017.12.022
    [14] C. Aouiti, F. Dridi, (μ,ν)-Pseudo-almost automorphic solutions for high-order Hopfield bidirectional associative memory neural networks, Neural Comput. Applic., 32 (2020), 1435–1456. https://doi.org/10.1007/s00521-018-3651-6 doi: 10.1007/s00521-018-3651-6
    [15] C. Aouiti, I. B. Gharbia, J. D. Cao, A. Alsaedi, Dynamics of impulsive neutral-type BAM neural networks, J. Franklin I., 356 (2019), 2294–2324. https://doi.org/10.1016/j.jfranklin.2019.01.028 doi: 10.1016/j.jfranklin.2019.01.028
    [16] K. Ding, N. J. Huang, Global robust exponential stability of interval general BAM neural network with delays, Neural Process. Lett., 23 (2006), 171–182. https://doi.org/10.1007/s11063-005-5090-5 doi: 10.1007/s11063-005-5090-5
    [17] C. J. Xu, P. L. Li, Y. C. Pang, Global exponential stability for interval general bidirectional associative memory (BAM) neural networks with proportional delays, Math. Method. Appl. Sci., 39 (2016), 5720–5731. https://doi.org/10.1002/mma.3957 doi: 10.1002/mma.3957
    [18] Z. Q. Zhang, W. B. Liu, D. M. Zhou, Global asymptotic stability to a generalized Cohen-Grossberg BAM neural networks of neutral type delays, Neural Networks, 25 (2012), 94–105. https://doi.org/10.1016/j.neunet.2011.07.006 doi: 10.1016/j.neunet.2011.07.006
    [19] D. S. Wang, L. H. Huang, Z. W. Cai, On the periodic dynamics of a general Cohen-Grossberg BAM neural networks via differential inclusions, Neurocomputing, 118 (2013), 203–214. https://doi.org/10.1016/j.neucom.2013.02.030 doi: 10.1016/j.neucom.2013.02.030
    [20] Z. Q. Zhang, K. Y. Liu, Existence and global exponential stability of a periodic solution to interval general bidirectional associative memory (BAM) neural networks with multiple delays on time scales, Neural Networks, 24 (2011), 427–439. https://doi.org/10.1016/j.neunet.2011.02.001 doi: 10.1016/j.neunet.2011.02.001
    [21] X. F. Li, D. Ding, J. Z. Feng, S. B. Hu, Existence and exponential stability of anti-periodic solutions for interval general bidirectional associative memory neural networks with multiple delays, Adv. Differ. Equ., 2016 (2016), 190. https://doi.org/10.1186/s13662-016-0882-7 doi: 10.1186/s13662-016-0882-7
    [22] L. Duan, Existence and global exponential stability of pseudo almost periodic solutions of a general delayed BAM neural networks, J. Syst. Sci. Complex., 31 (2018), 608–620. https://doi.org/10.1007/s11424-017-6180-y doi: 10.1007/s11424-017-6180-y
    [23] C. Aouiti, F. Dridi, New results on interval general Cohen-Grossberg BAM neural networks, J. Syst. Sci. Complex., 33 (2020), 944–967. https://doi.org/10.1007/s11424-020-8048-9 doi: 10.1007/s11424-020-8048-9
    [24] Y. S. Dong, Y. Han, T. T. Dai, Existence and exponential stability of almost periodic solutions to general BAM neural networks with leakage delays on time scales, Chinese Quarterly Journal of Mathematics, 37 (2022), 189–202. https://doi.org/10.13371/j.cnki.chin.q.j.m.2022.02.008 doi: 10.13371/j.cnki.chin.q.j.m.2022.02.008
    [25] Y. Li, L. Yang, W. Q. Wu, Anti-periodic solution for impulsive BAM neural networks with time-varying leakage delays on time scales, Neurocomputing, 149 (2015), 536–545. https://doi.org/10.1016/j.neucom.2014.08.020 doi: 10.1016/j.neucom.2014.08.020
    [26] S. H. Cai, Q. H. Zhang, Existence and stability of periodic solutions for impulsive fuzzy BAM Cohen-Grossberg neural networks on time scales, Adv. Differ. Equ., 2016 (2016), 64. https://doi.org/10.1186/s13662-016-0762-1 doi: 10.1186/s13662-016-0762-1
    [27] C. Aouiti, M. S. M'hamdi, J. D. Cao, A. Alsaedi, Piecewise pseudo almost periodic solution for impulsive generalised high-order Hopfield neural networks with leakage delays, Neural Process. Lett., 45 (2016), 615–648. https://doi.org/10.1007/s11063-016-9546-6 doi: 10.1007/s11063-016-9546-6
    [28] C. Wang, Piecewise pseudo-almost periodic solution for impulsive non-autonomous high-order Hopfield neural networks with variable delays, Neurocomputing, 171 (2016), 1291–1301. https://doi.org/10.1016/j.neucom.2015.07.054 doi: 10.1016/j.neucom.2015.07.054
    [29] C. Aouiti, E. A. Assali, Stability analysis for a class of impulsive bidirectional associative memory (BAM) neural networks with distributed delays and leakage time-varying delays, Neural Process. Lett., 50 (2019), 851–885. https://doi.org/10.1007/s11063-018-9937-y doi: 10.1007/s11063-018-9937-y
    [30] C. Aouiti, E. A. Assali, I. B. Gharbia, Y. E. Foutayeni, Existence and exponential stability of piecewise pseudo almost periodic solution of neutral-type inertial neural networks with mixed delay and impulsive perturbations, Neurocomputing, 357 (2019), 292–309. https://doi.org/10.1016/j.neucom.2019.04.077 doi: 10.1016/j.neucom.2019.04.077
    [31] C. Aouiti, I. B. Gharbia, Piecewise pseudo almost-periodic solutions of impulsive fuzzy cellular neural networks with mixed delays, Neural Process. Lett., 51 (2020), 1201–1225. https://doi.org/10.1007/s11063-019-10130-9 doi: 10.1007/s11063-019-10130-9
    [32] M. Abdelaziz, F. Cherif, Piecewise asymptotic almost periodic solutions for impulsive fuzzy Cohen-Grossberg neural networks, Chaos Soliton. Fract., 132 (2020), 109575. https://doi.org/10.1016/j.chaos.2019.109575 doi: 10.1016/j.chaos.2019.109575
    [33] M. Bohner, G. T. Stamov, I. M. Stamova, Almost periodic solutions of Cohen-Grossberg neural networks with time-varying delay and variable impulsive perturbations, Commun. Nonlinear Sci., 80 (2020), 104952. https://doi.org/10.1016/j.cnsns.2019.104952 doi: 10.1016/j.cnsns.2019.104952
    [34] A. M. Fink, Almost periodic differential equations, Heidelberg: Springer, 1974. http://doi.org/10.1007/BFb0070324
    [35] A. M. Samoilenko, N. A. Perestyuk, Impulsive differential equations, Singapore: World Scientific, 1995. https://doi.org/10.1142/2892
    [36] G. T. Stamov, Almost periodic solutions of impulsive differential equations, Heidelberg: Springer, 2012. http://doi.org/10.1007/978-3-642-27546-3
    [37] F. Cherif, Pseudo almost periodic solutions of impulsive differential equations with delay, Differ. Equ. Dyn. Syst., 22 (2014), 73–91. http://doi.org/10.1007/s12591-012-0156-0 doi: 10.1007/s12591-012-0156-0
    [38] J. W. Liu, C. Y. Zhang, Composition of piecewise pseudo almost periodic functions and applications to abstract impulsive differential equations, Adv. Differ. Equ., 2013 (2013), 11. https://doi.org/10.1186/1687-1847-2013-11 doi: 10.1186/1687-1847-2013-11
    [39] C. Aouiti, Oscillation of impulsive neutral delay generalized high-order Hopfield neural networks, Neural Comput. Applic., 29 (2018), 477–495. https://doi.org/10.1007/s00521-016-2558-3 doi: 10.1007/s00521-016-2558-3
    [40] Z. N. Xia, Pseudo almost periodic mild solution of nonautonomous impulsive integro-differential equations, Mediterr. J. Math., 13 (2016), 1065–1086. https://doi.org/10.1007/s00009-015-0532-4 doi: 10.1007/s00009-015-0532-4
    [41] C. Y. Zhang, Almost periodic type functions and ergodicity, Dordrecht: Springer, 2003.
  • Reader Comments
  • © 2023 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(2472) PDF downloads(73) Cited by(0)

Figures and Tables

Figures(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog