
The global exponential synchronization (GES) problem of a class of discrete-time high-order bidirectional associative memory neural networks (BAMNNs) with multiple time-varying delays (T-VDs) is studied. We investigate novel delay-dependent global exponential stability criteria for the error system by proposing a mathematical induction method. The global exponential stability criteria that have been obtained are described through linear scalar inequalities. These exponential synchronization conditions are very simple and convenient for verification based on standard software tools (such as YALMIP). Lastly, an instance is presented to demonstrate the validity of the theoretical findings.
Citation: Er-yong Cong, Li Zhu, Xian Zhang. Global exponential synchronization of discrete-time high-order BAM neural networks with multiple time-varying delays[J]. AIMS Mathematics, 2024, 9(12): 33632-33648. doi: 10.3934/math.20241605
[1] | 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. AIMS Mathematics, 2023, 8(9): 21828-21855. doi: 10.3934/math.20231113 |
[2] | 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 |
[3] | 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 |
[4] | Er-Yong Cong, Yantao Wang, Xian Zhang, Li Zhu . A new approach for accuracy-preassigned finite-time exponential synchronization of neutral-type Cohen–Grossberg memristive neural networks involving multiple time-varying leakage delays. AIMS Mathematics, 2025, 10(5): 10917-10942. doi: 10.3934/math.2025496 |
[5] | Călin-Adrian Popa . Synchronization of Clifford-valued neural networks with leakage, time-varying, and infinite distributed delays on time scales. AIMS Mathematics, 2024, 9(7): 18796-18823. doi: 10.3934/math.2024915 |
[6] | 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 |
[7] | Qinghua Zhou, Li Wan, Hongbo Fu, Qunjiao Zhang . Exponential stability of stochastic Hopfield neural network with mixed multiple delays. AIMS Mathematics, 2021, 6(4): 4142-4155. doi: 10.3934/math.2021245 |
[8] | 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 |
[9] | Dong Pan, Huizhen Qu . Finite-time boundary synchronization of space-time discretized stochastic fuzzy genetic regulatory networks with time delays. AIMS Mathematics, 2025, 10(2): 2163-2190. doi: 10.3934/math.2025101 |
[10] | N. Mohamed Thoiyab, Mostafa Fazly, R. Vadivel, Nallappan Gunasekaran . Stability analysis for bidirectional associative memory neural networks: A new global asymptotic approach. AIMS Mathematics, 2025, 10(2): 3910-3929. doi: 10.3934/math.2025182 |
The global exponential synchronization (GES) problem of a class of discrete-time high-order bidirectional associative memory neural networks (BAMNNs) with multiple time-varying delays (T-VDs) is studied. We investigate novel delay-dependent global exponential stability criteria for the error system by proposing a mathematical induction method. The global exponential stability criteria that have been obtained are described through linear scalar inequalities. These exponential synchronization conditions are very simple and convenient for verification based on standard software tools (such as YALMIP). Lastly, an instance is presented to demonstrate the validity of the theoretical findings.
In recent years, neural networks (NNs) have played an important role in fields such as addressable memory, pattern recognition, and optimization control due to their nonlinear computing capabilities and powerful parallel processing [1,2,3,4]. In 1986, Lee et al. [5] described a form of higher-order correlation formalism tensor for NNs, proposed high-order NNs based on ordinary low-order NNs. The model can simulate auto associative, heteroassociative, as well as multiassociative memory. In [6,7], several higher-order connection-weight models with different structures are proposed. Recently, an increasing number of scholars working on NN research have turned their attention from low-order systems to high-order systems, trying to overcome the limitations of traditional NNs. These higher-order models have higher fault tolerance, larger storage capacity, faster convergence speed, and stronger approximation properties than first-order NNs.
As one of the interconnected NNs, bidirectional associative memory NNs (BAMNNs) [8,9] consist of two-layers of heterogeneous associative circuits, which extend the functionality of single-layer neural networks and have the functions of memory and information association. For quite some time, most researchers have been focusing on low-order BAMNNs rather than high-order BAMNNs [10,11,12,13,14,15]. However, due to some shortcomings of low-order BAMNNs, it becomes very important to incorporate high-order interactions into such BAMNNs. Therefore, Simpson [7] put forward a class of higher-order delayed BAMNNs. In particular, second-order BAMNNs can enhance the storage capacity but require more connections. Since then, different architectures with high-order connections have been utilized to construct desirable BAMNNs [16,17,18].
It is widely known that synchronization is of utmost importance for BAMNNs. Identifying certain conditions to ensure the achievement of synchronization in drive-response systems is a significant research topic for BAMNNs. The term "synchronization" has a long history and denotes the state where two or more systems display common dynamic behaviors. When we study the dynamic behaviors of NNs, synchronization can play an extremely crucial role. In the past few years, the synchronization problem of various types of BAMNNs has drawn extensive attention due to their broad applications in multiple fields, including pattern recognition, associative memories, automatic control engineering, combinatorial optimization, signal processing, and parallel computation.
Up to now, it has been revealed that the research on the synchronization problem for delayed drive-response BAMNNs holds great significance for fundamental science. A large number of studies have emerged regarding the study of different types of synchronization problems for delayed BAMNNs. for example, long-time synchronization[19,20,21,22], finite-time synchronization [23,24,25] and fixed-time synchronization [26,27,28,29]. References [19] and [21] respectively studied the synchronization problems of single inertia BAMNN and delayed BAMNN by using matrix measure theory. They gave several criteria for GES by using Halanay inequality methods and Lyapunov stability theory, respectively. The proposed criteria are independent of delay parameters. Furthermore, valuable new insights into the stability and synchronization of BAMNNs are put forward. In reference [20], the authors concentrate impulsive synchronization of delayed memristive BAMNNs. By employing the so-called linear matrix inequality (LMI) approach, which is based on the time-varying Lyapunov function, the time-dependent impulsive results for the exponential stability of the error system are derived. In references [19,20,21], the Lyapunov functional method, LMI method, and matrix measure method are mainly applied to study synchronization problems. While reference [30] presents a new global synchronization research method and gives a criterion for global asymptotic synchronization of BAMNNs by using the integral inequality technique. This method and result extend the research on global synchronization of NNs. In reference [31], the general decay synchronization problem of BAMNNs with T-VD and distributed delay is studied by using Lyapunov method and useful inequality techniques. Sufficient criteria for the GDS of BAMNNs are given. In [32], sufficient conditions for global robust exponential synchronization of interval BAMNNs are given through a direct method based on the system solution. This method avoids the difficulty of establishing the Lyapunov–Krasovskii functional. Moreover, the derived global robust exponential synchronization criterion is simpler and easier to implement.
Regarding the achievement of exponential synchronization between dynamical systems, previous studies have employed various methods such as the matrix measure strategy and the method based on the finite-time stability theorem [24], the Lyapunov function method [33,34,35,36], the analytical method [37,38,39], the figure analysis method [40], and the integral inequality method [41,42,43,44]. However, there are few other methods apart from those mentioned above for studying the synchronization problem of delayed BAMNNs. This has motivated us to seek another method to investigate the GES problem of BAMNNs.
Furthermore, in computer simulation practices, discrete-time networks have certain advantages over continuous-time networks in transmitting digital information. Thus, researching the dynamic characteristics of discrete-time high-order NNs holds significant importance. However, aside from the aforementioned methods, there are very few other approaches for studying the synchronization problem of discrete-time delayed high-order BAMNNs. This spurs us to seek another method to investigate the GES problem of delayed discrete-time high-order BAMNNs. Based on the above discussion, we propose a novel method for synchronizing the drive-response system.
The primary objective of this paper is to design a control law that can globally exponentially synchronize the delayed discrete-time high-order BAMNNs. Additionally, it aims to obtain a time convergence that is more precise and has a high level of accuracy. The contributions of this article are:
(1) The synchronization criteria deduced by this approach are merely equivalent to solving a few straightforward linear scalar inequalities and are comparatively intuitive.
(2) The controller gains are in fact computed using the parameters of the discrete-time high-order BAMNN itself, and this can significantly decrease the computational complexity.
(3) The synchronization criterion derived is relatively straightforward and can be readily solved by utilizing YALMIP software.
The remaining portion is structured as follows: The elaboration and preparatory work of the problem will be conducted in the next section. The major achievements of this research, a new standard for GES, are presented in Section 3. Illustrative examples are provided in Section 4 to verify the validity of the obtained results. Finally, in Section 5, we present our conclusions.
Notations: The real number and integer set are denoted by R and Z, respectively. For positive integers l and s where l≤s, let [l,s]Z be the set that contains all the positive integers ranging from l to s. The symbol Rl×s denotes the set consisting of l×s matrices. Rl×s⪰ and Rl×s≻ are subsets of Rl×s, with the former containing all nonnegative matrices and the latter containing all positive matrices. In a similar vein, we also utilize R⪰, R≻, among others. When s→∞, the limit case of [l,s]Z is denoted as [l,∞)Z. For S=[sij]∈Rl×s and K=[kij]∈Rl×s, the matrix [sijkij], denoted by S∘K, refers to the Hadamard product of S and K, and the notation S⪰K (or S⪯K) denotes sij≥kij (or sij≤kij). If sij>kij (or sij<kij), we say S≻K (or S≺K). Let |S|=[|sij|]. Then |SK|⪯|S||K| for all S∈Rp×q and K∈Rq×r. The notation ‖⋅‖2 represents the Euclidean norm, and the notation ⊗ denotes the Kronecker product of matrices.
We consider the following discrete-time high-order BAMNNs with multiple T-VDs:
ζi(t+1)=aiζi(t)+m∑j=1[bijfj(ϑj(t))+cijfj(ϑj(t−δij(t)))]+m∑j=1m∑l=1dijlgj(ϑj(t−δijl(t)))gl(ϑl(t−δijl(t)))+Ii(t),i∈[1,n]Z,t∈[0,∞)Z, | (2.1a) |
ϑj(t+1)=ˆajϑj(t)+n∑i=1[ˆbji˜fi(ζi(t))+ˆcji˜fi(ζi(t−σji(t)))]+n∑i=1n∑r=1ˆdjir˜gi(ζi(t−σjir(t)))˜gr(ζr(t−σjir(t)))+˜Ij(t),j∈[1,m]Z,t∈[0,∞)Z, | (2.1b) |
where ai,ˆaj∈(−1,1); Constants bij, ˆbji, cij, ˆcji, dijl and ˆdjir represent the connection weights; ζi(t) and ϑj(t) denote the ith and jth neuronal states of layer-X and layer-Y, respectively; fj:R→[−s(1)j,s(1)j], gj:R→[−s(2)j,s(2)j], ˜fi:R→[−˜s(1)i,˜s(1)i] and ˜gi:R→[−˜s(2)i,˜s(2)i] denote the neuronal activation functions; s(1)j,s(2)j,˜s(1)i and ˜s(2)i are known positive constants; For the known positive integers ˉδij, ˉδijl, ˉσji and ˉσjir, the multiple T-VDs is represented as δij:[0,∞)Z→[0,ˉδij], δijl:[0,∞)Z→[0,ˉδijl], σji:[0,∞)Z→[0,ˉσji] and σjir:[0,∞)Z→[0,ˉσjir]; and Ii(t) and ˜Ij(t) are variable external input.
Remark 2.1. The BAMNN extends the single-layer auto-associative Hebbian correlator to a two-layer pattern matched hetero-associative circuits, and extracts the complete and clear patterns stored in memory from incomplete or fuzzy patterns. The network, as in (2.1), can store pairs of patterns or memories and search for them in both forward and backward directions. Therefore, the BAMNN exhibits relatively excellent information–memory and information–association capabilities.
We require these assumptions.
Assumption 2.1. For any α1,α2∈R subject to α1≠α2, there are ˜β(1)i, β(1)j∈R≻ such that ˜fi(0)=fj(0), 0≤˜fi(α1)−˜fi(α2)α1−α2≤˜β(1)i and 0≤fj(α1)−fj(α2)α1−α2≤β(1)j where ∀i∈[1,n]Z, ∀j∈[1,m]Z.
Assumption 2.2. For any α1,α2∈R subject to α1≠α2, there are ˜β(2)i,β(2)j∈R≻ such that ˜gi(0)=gj(0)=0, 0≤˜gi(α1)−˜gi(α2)α1−α2≤˜β(2)i, and 0≤gj(α1)−gj(α2)α1−α2≤β(2)j, ∀i∈[1,n]Z, ∀j∈[1,m]Z.
We regard discrete-time high-order BAMNN (2.1) as a master drive system and consider the following form of slave matching response system:
ˆζi(t+1)=aiˆζi(t)+m∑j=1[bijfj(ˆϑj(t))+cijfj(ˆϑj(t−δij(t)))]+m∑j=1m∑l=1dijlgj(ˆϑj(t−δijl(t)))gl(ˆϑl(t−δijl(t)))+Ii(t)+Ui(t),i∈[1,n]Z,t∈[0,∞)Z, | (2.2a) |
ˆϑj(t+1)=ˆajˆϑj(t)+n∑i=1[ˆbji˜fi(ˆζi(t))+ˆcji˜fi(ˆζi(t−σji(t)))]+n∑i=1n∑r=1ˆdjir˜gi(ˆζi(t−σjir(t)))˜gr(ˆζr(t−σjir(t)))+˜Ij(t)+Vj(t),j∈[1,m]Z,t∈[0,∞)Z, | (2.2b) |
where Ui(t) and Vj(t) are the controllers to realize the GES.
Let ηi(t)=ζi(t)−ˆζi(t), ˜ηj(t)=ϑj(t)−ˆϑj(t). Take ηi(t) and ˜ηj(t) as the synchronization error variables. Then, from Eqs (2.1) and (2.2), we can obtain that the error dynamical system:
ηi(t+1)=aiηi(t)+m∑j=1[bijf∗j(˜ηj(t))+cijf∗j(˜ηj(t−δij(t)))]+m∑j=1m∑l=1dijl[gj(ϑj(t−δijl(t)))g∗l(˜ηl(t−δijl(t)))+g∗j(˜ηj(t−δijl(t)))gl(ˆϑl(t−δijl(t)))]−Ui(t),i∈[1,n]Z,t∈[0,∞)Z, | (2.3a) |
˜ηj(t+1)=ˆaj˜ηj(t)+n∑i=1[ˆbji˜f∗i(ηi(t))+ˆcji˜f∗i(ηi(t−σji(t)))]+n∑i=1n∑r=1ˆdjir[˜gi(ζi(t−σjir(t)))˜g∗r(ηr(t−σjir(t)))+˜g∗i(ηi(t−σjir(t)))˜gr(ˆζr(t−σjir(t)))]−Vj(t),j∈[1,m]Z,t∈[0,∞)Z, | (2.3b) |
where
f∗j(˜ηj(⋅))=fj(˜ηj(⋅)+ˆϑj(⋅))−fj(ˆϑj(⋅)), |
g∗j(˜ηj(⋅))=gj(˜ηj(⋅)+ˆϑj(⋅))−gj(ˆϑj(⋅)), |
˜f∗i(ηi(⋅))=˜fi(ηi(⋅)+ˆζi(⋅))−˜fi(ˆζi(⋅)), |
˜g∗i(ηi(⋅))=˜gi(ηi(⋅)+ˆζi(⋅))−˜gi(ˆζi(⋅)). |
Due to Assumptions 2.1 and 2.2, we deduce that
|f∗j(ℏ)|≤β(1)j|ℏ|, |g∗j(ℏ)|≤β(2)j|ℏ|, |˜f∗i(ℏ)|≤˜β(1)i|ℏ|, |˜g∗i(ℏ)|≤˜β(2)i|ℏ|, ℏ∈R, i∈[1,n]Z,j∈[1,m]Z. | (2.4) |
Let ν=max1≤i,r≤n,1≤j,l≤mmax{ˉδij,ˉσji,ˉδijl,ˉσjir}. The symbol C([−ν,0]Z,Rn) denotes the set that consists of all functions φ:[−ν,0]Z→Rn. Let there be a norm ‖⋅‖ on Rn×Rm. It is defined such that for ‖(α,β)‖=(‖α‖22+‖β‖22)1/2, α∈Rn, β∈Rm. Similarly, a norm ‖(⋅,⋅)‖ν can be defined on C([−ν,0]Z,Rn)×C([−ν,0]Z,Rm) by
‖(ϕ,˜ϕ)‖ν=sups∈[−ν,0]Zmax{‖ϕ(s)‖2,‖˜ϕ(s)‖2}, ∀ϕ∈C([−ν,0]Z,Rn), ∀ ˜ϕ∈C([−ν,0]Z,Rm). |
Definition 2.1. [4] The response discrete-time high-order BAMNN (2.2) and the discrete-time high-order BAMNN (2.1) are said to be GES with a decay rate λ if there exist β,λ∈R≻ and controllers Ui(t) and Vj(t) such that arbitrary solution (η(t),˜η(t)) of error dynamical system (2.3) satisfying
‖(η(t),˜η(t))‖≤βe−λt‖(ϕ, ˜ϕ)‖ν,∀t∈[ν,∞)Z, |
where (ϕ, ˜ϕ)∈C([−ν,0]Z,Rn)×C([−ν,0]Z,Rm) is the initial functions corresponding to the solution (η(t),˜η(t)); η(t)=[η1(t) … ηn(t)]T and ˜η(t)=[˜η1(t) … ˜ηm(t)]T.
The objective of this paper is to design a state feedback controller of the form:
Ui(t)=(ρi−ai)ηi(t), i∈[1,n]Z,t∈[0,∞)Z,Vj(t)=(ˆρj−ˆaj)˜ηj(t),j∈[1,m]Z,t∈[0,∞)Z, | (2.5) |
which makes the discrete-time high-order BAMNNs (2.1) and (2.2) achieve GES, where ρi and ˆρj are the controller gains to be determined.
Remark 2.2. Discrete-time high-order BAMNN is a typical recurrent neural network, which plays an important role in pattern recognition and combinatorial optimization. At present, although many scholars have studied the dynamic characteristics of high-order BAMNN with time delays, most of them have focused on complex continuous-time network models, while the research on discrete-time network models is relatively scarce. Therefore, this paper conducts a study on the synchronization problem of a discrete-time high-order BAMNN model with multiple T-VDs. In addition, an important role is also played by BAMNN in the fields of signal processing and artificial intelligence.
Set
Ξβ=eβˉδ∘|C|Γ1+|B|Γ1+ℵ, eβˉδ=[eβˉδij], B=[bij], C=[cij],ℵ=GT(Eβ∘|D|+ˆEβ∘|ˆD|)Γ2,G=Im⊗P,P=[s(2)1s(2)2…s(2)m]T,D=[DT1DT2…DTm],ˆD=[D1D2…Dm],Di=[dijl],Eβ=[ETβ,1ETβ,2…ETβ,m],ˆEβ=[Eβ,1Eβ,2…Eβ,m],Eβ,i=[eβˉδijl],˜Ξβ=eβˉσ∘|˜C|˜Γ1+|˜B|˜Γ1+˜ℵ, eβˉσ=[eβˉσji], ˜B=[ˆbji], ˜C=[ˆcji],˜ℵ=HT(Fβ∘|˜D|+ˆFβ∘|ˆ˜D|)˜Γ2,H=In⊗Q,Q=[s(2)1s(2)2…s(2)n]T,˜D=[˜DT1˜DT2…˜DTn],ˆ˜D=[˜D1˜D2…˜Dn],˜Dj=[ˆdjil],Fβ=[FTβ,1FTβ,2…FTβ,n],ˆFβ=[Fβ,1Fβ,2…Fβ,n],Fβ,j=[eβˉσjil],Θβ,k=−e−βIk,Γ1=diag(β(1)1,…,β(1)m), ˜Γ1=diag(˜β(1)1,…,˜β(1)n),Γ2=diag(β(2)1,…,β(2)m), ˜Γ2=diag(˜β(2)1,…,˜β(2)n). |
Theorem 3.1. If there exists φ∈Rn≻, ψ∈Rm≻, ˜u∈Rn≻, ˜v∈Rm≻ and β∈R≻ such that
Θβ,n˜u+Ξβ˜v+φ⪯0, | (3.1) |
˜Ξβ˜u+Θβ,m˜v+ψ⪯0, | (3.2) |
then the error system (2.3) is globally exponentially stable, that is, the drive system discrete-time high-order BAMNN (2.1) and response system discrete-time high-order BAMNN (2.2) achieve GES via the controllers in (2.5), where ρi=±φi˜u−1i and ˆρj=±ψj˜v−1j, where ˜ui, ˜vj, φi and ψj are the ith and jth components of ˜u, ˜v, φ and ψ, respectively.
Proof. Choose Υ>0 such that
Υ˜u≻[1 ⋯ 1]T, Υ˜v≻[1 ⋯ 1]T. |
For any fixed ϕ∈C([−ν,0]Z,Rn) and ˜ϕ∈C([−ν,0]Z,Rm), define
ˆu(t)=Υ‖(ϕ,˜ϕ)‖νe−βt˜u, t∈[−ν,∞)Z, | (3.3) |
ˆv(t)=Υ‖(ϕ,˜ϕ)‖νe−βt˜v, t∈[−ν,∞)Z, | (3.4) |
where (ϕ,˜ϕ) serve as the initial functions.
Suppose (η(t),˜η(t)) is the solution of (2.3). Next, the following expression will be proved by using the mathematical induction method
|η(t)|⪯ˆu(t), |˜η(t)|⪯ˆv(t), t∈[−ν,∞)Z. | (3.5) |
Obviously, in combination with the definition of ‖⋅‖ν and the selection of Υ, we have
|η(k)|⪯ˆu(k), |˜η(k)|⪯ˆv(k), ∀k∈[−ν,0]Z. |
Suppose that for any fixed k≥0, when t≤k, the inequality (3.5) holds. For any i∈[1,n]Z, when t=k+1, using (2.3a), (2.5), (2.4), and Assumptions 2.1 and 2.2, we obtain
|ηi(k+1)|≤|ρi||ηi(k)|+m∑j=1[|bij||f∗j(˜ηj(k))|+|cij||f∗j(˜ηj(k−δij(k)))|]+m∑j=1m∑l=1|dijl|[|gj(ϑj(k−δijl(k)))||g∗l(˜ηl(k−δijl(k)))|+|g∗j(˜ηj(k−δijl(k)))||gl(ˆϑl(k−δijl(k)))|]≤|ρi||ηi(k)|+m∑j=1[|bij|β(1)j|˜ηj(k)|+|cij|β(1)j|˜ηj(k−δij(k))|]+m∑j=1m∑l=1|dijl|[|gj(ϑj(k−δijl(k)))|β(2)l|˜ηl(k−δijl(k))|+β(2)j|˜ηj(k−δijl(k))||gl(ˆϑl(k−δijl(k)))|]≤|ρi||ηi(k)|+m∑j=1[|bij|β(1)j|˜ηj(k)|+|cij|β(1)j|˜ηj(k−δij(k))|]+m∑j=1m∑l=1|dijl|[s(2)jβ(2)l|˜ηl(k−δijl(k))|+s(2)lβ(2)j|˜ηj(k−δijl(k))|]. |
Utilizing the inductive hypothesis, we are able to obtain
|ηi(k+1)|≤|ρi|ˆui(k)+m∑j=1[|bij|β(1)jˆvj(k)+|cij|β(1)jˆvj(k−δij(k))]+m∑j=1m∑l=1|dijl|[s(2)jβ(2)lˆvl(k−δijl(k))+s(2)lβ(2)jˆvj(k−δijl(k))]. | (3.6) |
By substituting Eqs (3.3) and (3.4) into (3.6), we obtain
|ηi(k+1)|≤|ρi|Υ‖(ϕ,˜ϕ)‖νe−βk˜ui+m∑j=1[|bij|β(1)jΥ‖(ϕ,˜ϕ)‖νe−βk˜vj+|cij|β(1)jΥ‖(ϕ,˜ϕ)‖νe−β(k−δij(k))˜vj]+m∑j=1m∑l=1|dijl|[s(2)jβ(2)lΥ‖(ϕ,˜ϕ)‖νe−β(k−δijl(k))˜vl+s(2)lβ(2)jΥ‖(ϕ,˜ϕ)‖νe−β(k−δijl(k))˜vj]≤Υ‖(ϕ,˜ϕ)‖νe−βk×{|ρi|˜ui+m∑j=1(|bij|β(1)j+|cij|β(1)jeβˉδij)˜vj+m∑j=1m∑l=1|dijl|[s(2)jβ(2)leβˉδijl˜vl+s(2)lβ(2)jeβˉδijl˜vj]}=Υ‖(ϕ,˜ϕ)‖νe−βk×{|ρi|˜ui+m∑j=1(|bij|β(1)j+|cij|β(1)jeβˉδij+m∑l=1|dijl|s(2)lβ(2)jeβˉδijl)˜vj+m∑j=1m∑l=1|dijl|s(2)jβ(2)leβˉδijl˜vl}=Υ‖(ϕ,˜ϕ)‖νe−βk×{|ρi|˜ui+m∑j=1(|bij|β(1)j+|cij|β(1)jeβˉδij+m∑l=1|dijl|s(2)lβ(2)jeβˉδijl)˜vj+m∑j=1m∑l=1|dilj|s(2)lβ(2)jeβˉδilj˜vj}=Υ‖(ϕ,˜ϕ)‖νe−βk×{|ρi|˜ui+m∑j=1[|bij|β(1)j+|cij|β(1)jeβˉδij+m∑l=1|dijl|s(2)lβ(2)jeβˉδijl+m∑l=1|dilj|s(2)lβ(2)jeβˉδilj]˜vj}≤Υ‖(ϕ,˜ϕ)‖νe−βk×{|ρi|˜ui+m∑j=1[|bij|β(1)j+|cij|β(1)jeβˉδij+m∑l=1s(2)l(|dijl|eβˉδijl+|dilj|eβˉδilj)β(2)j]˜vj}. | (3.7) |
Note that φi=|ρi|˜ui. Based on the arbitrariness of i∈[1,n]Z, we can conclude that (3.7) is equivalent to
|η(k+1)|⪯Υ‖(ϕ,˜ϕ)‖νe−βk(φ+Ξβ˜v). |
By making use of (3.1) and (3.3), we obtain
|η(k+1)|⪯Υ‖(ϕ,˜ϕ)‖νe−β(k+1)˜u=ˆu(k+1). | (3.8) |
Similarly, through a procedure similar to the one used in deriving (3.8), it is straightforward to that
|˜η(k+1)|⪯Υ‖(ϕ,˜ϕ)‖νe−β(k+1)˜v=ˆv(k+1). |
Therefore, (3.5) is true.
Then, in combination with Eqs (3.3)–(3.5), we obtain
‖(η(t),˜η(t))‖=(‖η(t)‖22+‖˜η(t)‖22)12≤(‖ˆu(t)‖22+‖ˆv(t)‖22)12=Υe−βt‖(ϕ,˜ϕ)‖ν(‖˜u‖22+‖˜v‖22)12, ∀t∈[0,∞)Z. |
Let μ=Υ(‖˜u‖22+‖˜v‖22)12. Then
‖(η(t),˜η(t))‖≤μe−βt‖(ϕ,˜ϕ)‖ν, ∀t∈[0,∞)Z. |
The arbitrariness of ϕ∈C([−ν,0]Z,Rn) and ˜ϕ∈C([−ν,0]Z,Rm) guarantees GES of the error system (2.3), that is, the discrete-time high-order BAMNN (2.1) and (2.2) achieve GES via the controllers in (2.5).
Next, the effectiveness of the results given in this paper will be illustrated through a specific numerical example.
Example 4.1. For n=2,m=2, consider the drive discrete-time high-order BAMNNs (2.1) and the response discrete-time high-order BAMNNs (2.2) with the controllers in (2.5), and the parameters are as below: a1=0.3,a2=0.4, ˆa1=0.5, ˆa2=0.6, b11=−1.6, b12=0.1, b21=−4.1, b22=3.2, ˆb11=1.1, ˆb12=0.5, ˆb21=−0.1, ˆb22=−2.1, c11=−1.4, c12=0.1, c21=0.2, c22=−2.3, ˆc11=−3, ˆc12=−2.5, ˆc21=0.3, ˆc22=−1.2, d111=5, d112=1, d121=0, d122=5, d211=2, d212=0, d221=0, d222=2, ˆd111=4, ˆd112=0, ˆd121=0, ˆd122=5, ˆd211=1, ˆd212=1, ˆd221=0, ˆd222=0, and
f1(s)=f2(s)=˜f1(s)=˜f2(s)=0.04tanh(s),g1(s)=g2(s)=˜g1(s)=˜g2(s)=0.4tanh(s),s∈R,σji(t)=pji+qjicos(tπ), δij(t)=rij+sijsin(tπ/2), i,j∈[1,2]Z,t∈[0,∞)Z, |
where r11=r12=r21=r22=10, s11=s12=s21=s22=10, p11=p12=p21=p22=10, q11=q21=q12=q22=10, δ111=δ112=δ121=δ122=δ211=δ212=δ221=δ222=3, σ111=σ112=σ121=σ122=σ211=σ212=σ221=σ222=3.
Clearly, ˉσ11=ˉσ12=ˉσ21=ˉσ22=ˉδ11=ˉδ12=ˉδ21=ˉδ22=20. Furthermore, when β(1)1=β(1)2=0.04, ˜β(1)1=˜β(1)2=0.04, β(2)1=β(2)2=0.4, and ˜β(2)1=˜β(2)2=0.4, Assumptions 2.1 and 2.2 are satisfied.
By solving the inequalities (3.1) and (3.2) in Theorem 3.1, the following feasible solutions are obtained:
˜u=[7.6343,5.6956]T,˜v=[11.0709,3.1479]T, |
φ=[5.9×103,4.6×103]T,ψ=[5.3×103,3.7×103]T. |
Consequently, the controller gains of the desired state feedback controllers are as follows:
ρ1=0.7709×10−3, ρ2=0.8025×10−3,ˆρ1=0.5×10−3, ˆρ2=1.2×10−3. |
It can be readily verified that the conditions of Theorem 3.1 in our paper are met. Consequently, based on Theorem 3.1 in our paper, the drive system given by (2.1) and the response system given by (2.2) are GES under the controllers (2.5). We choose the initial values of the state variables as ζ(s)=[0.6428 −0.1106]T, ϑ(s)=[0.2309 0.5839]T, ˆζ(s)=[1.6873 0.9528]T and ˆϑ(s)=[−1.2949 −0.3772]T, s∈[−20,0]Z. We also define the external input I1=I2=0,J1=J2=0. The error curves of drive–responses system η1(t),η2(t),˜η1(t) and ˜η2(t) are shown in Figures 5 and 6, the curves of variables ζ1(t),ζ2(t),ϑ1(t),ϑ2(t), ˆζ1(t),ˆζ2(t),ˆϑ1(t),ˆϑ2(t), are shown in Figures 1–4. It is readily observable that the error states are rapidly converging to the equilibrium point of zero. Therefore, it can be concluded that, in accordance with Theorem 3.1, the considered discrete-time high-order BAMNNs model (2.1) can achieve GES with system (2.2).
This paper addresses the issue of GES for discrete-time high-order BAMNNs of multiple T-VD. Utilizing the definition of GES, we initially derive delay-dependent GES criteria for the error dynamical system. Subsequently, a controller gain is provided. Finally, we present illustrative examples to demonstrate the applicability of the conclusions. Compared with the previous research results, the proposed method has the following advantages:
(1) The method directly employs the definition of GES and avoids the construction of any Lyapunov-Krasovskii function;
(2) The obtained sufficient conditions are in the form of linear scalar inequalities, which are easy to solve;
(3) With a small modification, the proposed method is applicable to more general NNs models.
The stability of the error system considered in this paper is the stability in the Lyapunov sense. What is mainly considered is the long-term behavior of the neural network, and it cannot be directly applied to the research on the synchronization in finite time. Therefore, it is necessary to propose more effective analysis and design methods.
Er-yong Cong: Writing-review & editing, writing-original draft, validation, investigation, conceptualization; Li Zhu: Writing-review & editing, visualization, software, investigation, funding acquisition; Xian Zhang: Writing-review & editing, resources, methodology, conceptualization. All authors have read and approved the final version of the manuscript for publication.
This work was supported in part by the Natural Science Foundation of Heilongjiang Province (No. LH2022F046 and No. YQ2022F015), the Project for Research Foundation of Young Doctor of Harbin University (No. HUDF2022112) and Harbin Science and Technology Plan Project (No. 2023ZCZJCG045 and No. 2023ZCZJCG046).
The authors would like to thank the anonymous reviewers for their helpful comments and suggestions, which greatly improves the original version of the paper.
All authors declare no conflicts of interest in this paper.
[1] |
J. Townsend, T. Chaton, J. M. Monteiro, Extracting relational explanations from deep neural networks: A survey from a neural-symbolic, IEEE Trans. Neural Netw. Learn. Syst., 31 (2020), 3456–3470. http://dx.doi.org/10.1109/TNNLS.2019.2944672 doi: 10.1109/TNNLS.2019.2944672
![]() |
[2] |
Y. Chen, X. Zhang, Y. Xue, Global exponential synchronization of high-order quaternion Hopfield neural networks with unbounded distributed delays and time-varying discrete delays, Math. Comput. Simul., 193 (2022), 173–189. http://dx.doi.org/10.1016/j.matcom.2021.10.012 doi: 10.1016/j.matcom.2021.10.012
![]() |
[3] |
S. K. Thangarajan, A. Chokkalingam, Integration of optimized neural network and convolutional neural network for automated brain tumor detection, Sensor Rev., 41 (2021), 16–34. http://dx.doi.org/10.1108/SR-02-2020-0039 doi: 10.1108/SR-02-2020-0039
![]() |
[4] |
Z. Dong, X. Wang, X. Zhang, M. Hu, T. N. Dinh, Global exponential synchronization of discrete-time high-order switched neural networks and its application to multi-channel audio encryption, Nonlinear Anal. Hybrid Syst., 47 (2023), 101291. http://dx.doi.org/10.1016/j.nahs.2022.101291 doi: 10.1016/j.nahs.2022.101291
![]() |
[5] |
Y. Lee, G. Doolen, H. H. Chen, G. Z. Sun, T. Maxwell, H. Y. Lee, et. al., Machine learning using a higher order correlation network, Phys. D, 22 (1986), 276–306. http://dx.doi.org/10.1016/0167-2789(86)90300-6 doi: 10.1016/0167-2789(86)90300-6
![]() |
[6] |
D. Psaltis, C. H. Park, J. Hong, Higher order associative memories and their optical implementations, Neural Netw., 1 (1988), 149–163. http://dx.doi.org/10.1016/0893-6080(88)90017-2 doi: 10.1016/0893-6080(88)90017-2
![]() |
[7] |
P. K. Simpson, Higher-ordered and intraconnected bidirectional associative memories, IEEE Trans. Syst. Man Cybernet Syst., 20 (1990), 637–653. http://dx.doi.org/10.1109/21.57276 doi: 10.1109/21.57276
![]() |
[8] |
B. Kosko, Adaptive bidirectional associative memories, Appl. Opt., 26 (1987), 4947–4960. http://dx.doi.org/10.1364/ao.26.004947 doi: 10.1364/ao.26.004947
![]() |
[9] |
B. Kosko, Bidirectional associative memories, IEEE Trans. Syst. Man Cybernet., 18 (1988), 49–60. http://dx.doi.org/10.1109/21.87054 doi: 10.1109/21.87054
![]() |
[10] |
X. Lou, B. Cui, On the global robust asymptotic stability of BAM neural networks with time-varying delays, Neurocomputing, 70 (2006), 273–279. http://dx.doi.org/10.1016/j.neucom.2006.02.020 doi: 10.1016/j.neucom.2006.02.020
![]() |
[11] |
X. Li, Exponential stability of Cohen–Grossberg-type BAM neural networks with time-varying delays via impulsive control, Neurocomputing, 73 (2009), 525–530. http://dx.doi.org/10.1016/j.neucom.2009.04.022 doi: 10.1016/j.neucom.2009.04.022
![]() |
[12] |
Z. Zhang, K. Liu, Y. Yang, New LMI-based condition on global asymptotic stability concerning BAM neural networks of neutral type, Neurocomputing, 81 (2012), 24–32. http://dx.doi.org/10.1016/j.neucom.2011.10.006 doi: 10.1016/j.neucom.2011.10.006
![]() |
[13] |
E. Y. Cong, X. Han, X. Zhang, Global exponential stability analysis of discrete-time BAM neural networks with delays: A mathematical induction approach, Neurocomputing, 379 (2020), 227–235. http://dx.doi.org/10.1016/j.neucom.2019.10.089 doi: 10.1016/j.neucom.2019.10.089
![]() |
[14] |
E. Y. Cong, X. Han, X. Zhang, New stabilization method for delayed discrete-time Cohen–Grossberg BAM neural networks, IEEE Access, 8 (2020), 99327–99336. http://dx.doi.org/10.1109/ACCESS.2020.2997905 doi: 10.1109/ACCESS.2020.2997905
![]() |
[15] |
L. Zhu, E. Y. Cong, X. Zhang, State estimation for a class of discrete-time BAM neural networks with multiple time-varying delays, IEEE Access, 11 (2023), 29314–29322. http://dx.doi.org/10.1109/ACCESS.2023.3260619 doi: 10.1109/ACCESS.2023.3260619
![]() |
[16] |
J. Cao, J. Liang, J. Lam, Exponential stability of high-order bidirectional associative memory neural networks with time delays, Phys. D, 199 (2004), 425–436. http://dx.doi.org/10.1016/j.physd.2004.09.012 doi: 10.1016/j.physd.2004.09.012
![]() |
[17] |
F. Wang, M. Liu, Global exponential stability of high-order bidirectional associative memory (BAM) neural networks with time delays in leakage terms, Neurocomputing, 177 (2016), 515–528. http://dx.doi.org/10.1016/j.neucom.2015.11.052 doi: 10.1016/j.neucom.2015.11.052
![]() |
[18] |
C. Aouiti, X. Li, F. Miaadi, A new LMI approach to finite and fixed time stabilization of high-order class of BAM neural networks with time-varying delays, Neural Process. Lett., 50 (2019), 815–838. http://dx.doi.org/10.1007/s11063-018-9939-9 doi: 10.1007/s11063-018-9939-9
![]() |
[19] |
J. D. Cao, Y. Wan, Matrix measure strategies for stability and synchronization of inertial BAM neural network with time delays, Neural Netw., 53 (2014), 165–172. http://dx.doi.org/10.1016/j.neunet.2014.02.003 doi: 10.1016/j.neunet.2014.02.003
![]() |
[20] |
K. Mathiyalagan, J. H. Park, R. Sakthivel, Synchronization for delayed memristive BAM neural networks using impulsive control with random nonlinearities, Appl. Math. Comput., 259 (2015), 967–979. http://dx.doi.org/10.1016/j.amc.2015.03.022 doi: 10.1016/j.amc.2015.03.022
![]() |
[21] |
Y. Li, C. Li, Matrix measure strategies for stabilization and synchronization of delayed BAM neural networks, Nonlinear Dyn., 84 (2016), 1759–1770. http://dx.doi.org/10.1007/s11071-016-2603-x doi: 10.1007/s11071-016-2603-x
![]() |
[22] |
H. Shen, Z. Huang, Z. Wu, J. Cao, J. H. Park, Nonfragile H∞ synchronization of BAM inertial neural networks subject to persistent dwell-time switching regularity, IEEE Trans. Cybernet., 52 (2021), 6591–6602. http://dx.doi.org/10.1109/TCYB.2021.3119199 doi: 10.1109/TCYB.2021.3119199
![]() |
[23] |
D. Chen, Z. Zhang, Finite-time synchronization for delayed BAM neural networks by the approach of the same structural functions, Chaos Soliton. Fract., 164 (2022), 112655. http://dx.doi.org/10.1016/j.chaos.2022.112655 doi: 10.1016/j.chaos.2022.112655
![]() |
[24] |
Y. Zhang, L. Li, H. Peng, J. Xiao, Y. Yang, M. Zheng, et al., Finite-time synchronization for memristor-based BAM neural networks with stochastic perturbations and time-varying delays, Internat. J. Robust Nonlinear Control, 28 (2018), 5118–5139. http://dx.doi.org/10.1002/rnc.4302 doi: 10.1002/rnc.4302
![]() |
[25] |
Z. Yang, Z. Zhang, Finite-time synchronization analysis for BAM neural networks with time-varying delays by applying the maximum-value approach with new inequalities, Mathematics, 10 (2022), 835. http://dx.doi.org/10.3390/math10050835 doi: 10.3390/math10050835
![]() |
[26] |
C. Chen, L. Li, H. Peng, Y. Yang, Fixed-time synchronization of memristor-based BAM neural networks with time-varying discrete delay, Neural Netw., 96 (2017), 47–54. http://dx.doi.org/10.1016/j.neunet.2017.08.012 doi: 10.1016/j.neunet.2017.08.012
![]() |
[27] |
H. Yan, Y. Qiao, J. Miao, Z. Ren, L. Duan, Fixed-time synchronization of delayed BAM neural networks via new fixed-time stability results and non-chattering quantized controls, J. Franklin Inst., 360 (2023), 10251–10274. http://dx.doi.org/10.1016/j.jfranklin.2023.07.044 doi: 10.1016/j.jfranklin.2023.07.044
![]() |
[28] |
J. Yang, G. Chen, S. Zhu, S. Wen, J. Hu, Fixed/prescribed-time synchronization of BAM memristive neural networks with time-varying delays via convex analysis, Neural Netw., 163 (2023), 53–63. http://dx.doi.org/10.1016/j.neunet.2023.03.031 doi: 10.1016/j.neunet.2023.03.031
![]() |
[29] |
Y. Cheng, H. Zhang, I. Stamova, J. Cao, Estimate scheme for fractional order-dependent fixed-time synchronization on caputo quaternion-valued BAM network systems with time-varying delays, J. Franklin Inst., 360 (2023), 2379–2403. http://dx.doi.org/10.1016/j.jfranklin.2022.10.055 doi: 10.1016/j.jfranklin.2022.10.055
![]() |
[30] |
F. Lin, Z. Zhang, Global asymptotic synchronization of a class of BAM neural networks with time delays via integrating inequality techniques, J. Syst. Sci. Complex., 33 (2020), 366–382. http://dx.doi.org/10.1007/s11424-019-8121-4 doi: 10.1007/s11424-019-8121-4
![]() |
[31] |
A. Muhammadhaji, Z. Teng, Synchronization stability on the BAM neural networks with mixed time delays, Int. J. Nonlinear Sci. Numer. Simul., 22 (2021), 99–109. http://dx.doi.org/10.1515/ijnsns-2019-0308 doi: 10.1515/ijnsns-2019-0308
![]() |
[32] |
J. Lan, X. Wang, X. Zhang, Global robust exponential synchronization of interval BAM neural networks with multiple time-varying delays, Circuits Syst. Signal Process., 43 (2024), 2147–2170. http://dx.doi.org/10.1007/s00034-023-02584-z doi: 10.1007/s00034-023-02584-z
![]() |
[33] |
L. Li, R. Xu, Q. Gan, J. Lin, A switching control for finite-time synchronization of memristor-based BAM neural networks with stochastic disturbances, Nonlinear Anal. Model. Control, 25 (2020), 958–979. http://dx.doi.org/10.15388/namc.2020.25.20557 doi: 10.15388/namc.2020.25.20557
![]() |
[34] |
R. Tang, X. Yang, X. Wan, Y. Zou, Z. Cheng, H. M. Fardoun, Finite-time synchronization of nonidentical BAM discontinuous fuzzy neural networks with delays and impulsive effects via non-chattering quantized control, Commun. Nonlinear Sci. Numer. Simul., 78 (2019), 104893. http://dx.doi.org/10.1016/j.cnsns.2019.104893 doi: 10.1016/j.cnsns.2019.104893
![]() |
[35] |
Z. Wang, X. Zhang, J. Qiao, H. Wu, T. Huang, Fuzzy fault-tolerant boundary control for nonlinear DPSs with multiple delays and stochastic actuator failures, IEEE Trans. Fuzzy Syst., 32 (2024), 3121–3131. https://doi.org/10.1109/TFUZZ.2024.3367870 doi: 10.1109/TFUZZ.2024.3367870
![]() |
[36] |
Z. Wang, B. Chen, J. Qiao, H. Wu, T. Huang, Fuzzy boundary sampled-data control for nonlinear DPSs with random time-varying delays, IEEE Trans. Fuzzy Syst., 32 (2024), 5872–5885. http://dx.doi.org/10.1109/TFUZZ.2024.3432795 doi: 10.1109/TFUZZ.2024.3432795
![]() |
[37] |
L. Zhang, Y. Yang, Finite time impulsive synchronization of fractional order memristive BAM neural networks, Neurocomputing, 384 (2020), 213–224. http://dx.doi.org/10.1016/j.neucom.2019.12.056 doi: 10.1016/j.neucom.2019.12.056
![]() |
[38] |
H. Li, J. Cao, C. Hua, L. Zhang, H. Jiang, Adaptive control-based synchronization of discrete-time fractional-order fuzzy neural networks with time-varying delays, Neural Netw., 168 (2023), 59–73. http://dx.doi.org/10.1016/j.neunet.2023.09.019 doi: 10.1016/j.neunet.2023.09.019
![]() |
[39] |
H. Li, J. Cao, C. Hu, H. Jiang, F. E. Alsaadi, Synchronization analysis of discrete-time fractional-order quaternion-valued uncertain neural networks, IEEE Trans. Neural Netw. Learn. Syst., 35 (2024), 14178–14189. http://dx.doi.org/10.1109/TNNLS.2023.3274959 doi: 10.1109/TNNLS.2023.3274959
![]() |
[40] |
Z. Zhou, Z. Zhang, M. Chen, Finite-time synchronization for fuzzy delayed neutral-type inertial BAM neural networks via the figure analysis approach, Int. J. Fuzzy Syst., 24 (2022), 229–246. http://dx.doi.org/10.1007/s40815-021-01132-8 doi: 10.1007/s40815-021-01132-8
![]() |
[41] |
M. Sader, A. Abdurahman, H. Jiang, General decay synchronization of delayed BAM neural networks via nonlinear feedback control, Appl. Math. Comput., 337 (2018), 302–314. http://dx.doi.org/10.1016/j.amc.2018.05.046 doi: 10.1016/j.amc.2018.05.046
![]() |
[42] |
Z. Zhang, J. Cao, Novel finite-time synchronization criteria for inertial neural networks with time delays via integral inequality method, IEEE Trans. Neural Netw. Learn. Syst., 30 (2018), 1476–1485. http://dx.doi.org/10.1109/TNNLS.2018.2868800 doi: 10.1109/TNNLS.2018.2868800
![]() |
[43] |
Z. Zhang, M. Chen, A. Li, Further study on finite-time synchronization for delayed inertial neural networks via inequality skills, Neurocomputing, 373 (2020), 15–23. http://dx.doi.org/10.1016/j.neucom.2019.09.034 doi: 10.1016/j.neucom.2019.09.034
![]() |
[44] |
Z. Zhang, A. Li, S. Yu, Finite-time synchronization for delayed complex-valued neural networks via integrating inequality method, Neurocomputing, 318 (2018), 248–260. http://dx.doi.org/10.1016/j.neucom.2018.08.063 doi: 10.1016/j.neucom.2018.08.063
![]() |