Research article

Further study on Hopf bifurcation and hybrid control strategy in BAM neural networks concerning time delay

  • Received: 24 January 2024 Revised: 26 February 2024 Accepted: 01 March 2024 Published: 10 April 2024
  • MSC : 34C23, 34K18, 37GK15, 39A11, 92B20

  • Delayed dynamical system plays a vital role in describing the dynamical phenomenon of neural networks. In this article, we proposed a class of new BAM neural networks involving time delay. The traits of solution and bifurcation behavior of the established BAM neural networks involving time delay were probed into. First, the existence and uniqueness is discussed using a fixed point theorem. Second, the boundedness of solution of the formulated BAM neural networks involving time delay was analyzed by applying an appropriate function and inequality techniques. Third, the stability peculiarity and bifurcation behavior of the addressed delayed BAM neural networks were investigated. Fourth, Hopf bifurcation control theme of the formulated delayed BAM neural networks was explored by virtue of a hybrid controller. By adjusting the parameters of the controller, we could control the stability domain and Hopf bifurcation onset, which was in favor of balancing the states of different neurons in engineering. To verify the correctness of gained major outcomes, computer simulations were performed. The acquired outcomes of this article were new and own enormous theoretical meaning in designing and dominating neural networks.

    Citation: 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[J]. AIMS Mathematics, 2024, 9(5): 13265-13290. doi: 10.3934/math.2024647

    Related Papers:

    [1] 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
    [2] 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
    [3] Yingyan Zhao, Changjin Xu, Yiya Xu, Jinting Lin, Yicheng Pang, Zixin Liu, Jianwei Shen . Mathematical exploration on control of bifurcation for a 3D predator-prey model with delay. AIMS Mathematics, 2024, 9(11): 29883-29915. doi: 10.3934/math.20241445
    [4] 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
    [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] Tian Xu, Ailong Wu . Stabilization of nonlinear hybrid stochastic time-delay neural networks with Lévy noise using discrete-time feedback control. AIMS Mathematics, 2024, 9(10): 27080-27101. doi: 10.3934/math.20241317
    [7] Changjin Xu, Chaouki Aouiti, Zixin Liu, Qiwen Qin, Lingyun Yao . Bifurcation control strategy for a fractional-order delayed financial crises contagions model. AIMS Mathematics, 2022, 7(2): 2102-2122. doi: 10.3934/math.2022120
    [8] Erxi Zhu . Time-delayed feedback control for chaotic systems with coexisting attractors. AIMS Mathematics, 2024, 9(1): 1088-1102. doi: 10.3934/math.2024053
    [9] Hairong Li, Yanling Tian, Ting Huang, Pinghua Yang . Hopf bifurcation and hybrid control of a delayed diffusive semi-ratio-dependent predator-prey model. AIMS Mathematics, 2024, 9(10): 29608-29632. doi: 10.3934/math.20241434
    [10] Yuehong Zhang, Zhiying Li, Wangdong Jiang, Wei Liu . The stability of anti-periodic solutions for fractional-order inertial BAM neural networks with time-delays. AIMS Mathematics, 2023, 8(3): 6176-6190. doi: 10.3934/math.2023312
  • Delayed dynamical system plays a vital role in describing the dynamical phenomenon of neural networks. In this article, we proposed a class of new BAM neural networks involving time delay. The traits of solution and bifurcation behavior of the established BAM neural networks involving time delay were probed into. First, the existence and uniqueness is discussed using a fixed point theorem. Second, the boundedness of solution of the formulated BAM neural networks involving time delay was analyzed by applying an appropriate function and inequality techniques. Third, the stability peculiarity and bifurcation behavior of the addressed delayed BAM neural networks were investigated. Fourth, Hopf bifurcation control theme of the formulated delayed BAM neural networks was explored by virtue of a hybrid controller. By adjusting the parameters of the controller, we could control the stability domain and Hopf bifurcation onset, which was in favor of balancing the states of different neurons in engineering. To verify the correctness of gained major outcomes, computer simulations were performed. The acquired outcomes of this article were new and own enormous theoretical meaning in designing and dominating neural networks.



    We all know that neural networks have displayed great application prospect in massive fields such as associative memory, combinatorial optimization, artificial intelligence, pattern recognition, signal processing and disease diagnosis [1,2,3,4,5]. In order to make fully use of neural networks to serve mankind, it is important for us to explore the dynamic properties of neural networks. In particular, time delay usually appears in neural networks due to the delay of signal propagation among different neurons. Thus, delayed neural networks are more reasonable to describe the interaction of real neurons. During the past several decades, many researchers pay great attention to the dynamics of delayed neural networks. Due to the existence of time delay, many dynamical behaviors of neural networks will change. In many cases, neural networks will lose their stability, produce periodic vibration, give rise to Hopf bifurcation, bring about chaotic phenomenon, and so on [6,7,8,9,10,11]. During the past decades, a number of outstanding achievements on neural networks with delays have been acquired. For instance, Abdurahman and Jiang [11] dealt with the general decay projective synchronization of memristor-based BAM neural networks with delays via nonlinear control technique. Kong et al. [12] explored the stability of almost periodic solutions to discontinuous BAM neural networks involving hybrid time-varying delays and D operator. Xu et al. [13] discussed the bifurcation issue of a simplified tri-neuron BAM networks with both time delays in frequency domain. Pratap et al. [14] made a detailed analysis on quasi-pinning synchronization and stabilization in a class of fractional-order BAM neural networks involving delays and discontinuous neuron activations. Maharajan et al. [15] set up a new global robust exponential stability condition for uncertain inertial-type BAM neural networks with discrete and distributed delays. For details, one can see [16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38].

    Bidirectional associative memory (BAM) neural networks, which were first proposed by Kosto [11], are a class of special neural networks. They include neurons in two fields and are able to store bipolar vector pairs [12]. In 2008, Sprott [13] proposed the following 4D BAM neural networks:

    {˙u1(t)=tanh[u4(t)u2(t)]au1(t),˙u2(t)=tanh[u1(t)+u4(t)]au2(t),˙u3(t)=tanh[u1(t)+u2(t)u4(t)]au3(t),˙u4(t)=tanh[u3(t)u2(t)]au4(t), (1.1)

    where ui(i=1,2,3,4) denote the state of the i neuron, a is a positive constant and the formula -aui has a relation with frictional damping and guarantees the boundedness of the solutions. Sprott [13] had studied the chaotic phenomenon of the 4D BAM neural networks (1.1). Noticing that different self-feedback time delays in the first neuron and the second neuron often exist in system (1.1), Vaishwar and Yadav [14] modified system (1.1) as the following 4D BAM neural networks concerning both inconsistent time delays:

    {˙u1(t)=tanh[u4(t)u2(t)]au1(tθ1),˙u2(t)=tanh[u1(t)+u4(t)]au2(tθ2),˙u3(t)=tanh[u1(t)+u2(t)u4(t)]au3(t),˙u4(t)=tanh[u3(t)u2(t)]au4(t), (1.2)

    where θ1>0,θ2>0 stand for two delays. Vaishwar and Yadav [14] investigated the stability behavior and Hopf bifurcation phenomenon of neural network model (1.2). Here, we would like to point out that all the neurons have self-feedback time delays. That is to say, the first neuron u1, the second neuron u2, the third neuron u3, and the fourth neuron u4 have different self-feedback time delays. For simplicity, we assume that self-feedback time delays for four neurons are same, then relying on the model (1.2), we can formulate the following 4D BAM neural networks concerning time delay:

    {˙u1(t)=tanh[u4(t)u2(t)]au1(tθ),˙u2(t)=tanh[u1(t)+u4(t)]au2(tθ),˙u3(t)=tanh[u1(t)+u2(t)u4(t)]au3(tθ),˙u4(t)=tanh[u3(t)u2(t)]au4(tθ), (1.3)

    where θ stands for a delay.

    Compared with the model (1.2), there exist more self-feedback delays (for neurons u3 and neuron u4) in model (1.2). Usually, delay will lead to the vary of dynamical behavior of neural networks. However, Vaishwar and Yadav [14] did not consider this problem. This motivates us to explore this model (1.2). In addition, delay-driven Hopf bifurcation plays a significant role in describing the interaction law in nonlinear delayed dynamical systems [14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35]. In neural network area, delay-driven Hopf bifurcation can effectively describe the transformation relationship of the state of different neurons. Thus we think that it is an important thing to deal with the delay-Hopf bifurcation in many kinds of neural networks. Stimulated by the viewpoint above, we are about to handle the delay-Hopf bifurcation and the Hopf bifurcation control problem for system (1.3). In particular, we handle the following topics: (Ⅰ) Prove the existence, uniqueness and non-negativeness of the solution to system (1.3). (Ⅱ) Investigate the stability behavior and the emergence Hopf bifurcation of system (1.3). (Ⅲ) Dominate Hopf bifurcation of system (1.3) by virtue of hybrid controller.

    The main highlights of this manuscript are stated as follows: (Ⅰ) Based on the previous studies, novel 4D BAM neural networks concerning time delay are formulated. (Ⅱ) A new delay-independent criterion on the stability and the appearance of bifurcation of system (1.3) is acquired. (Ⅲ) Taking advantage of hybrid controller, the stability domain and the time of the emergence of bifurcation of system (1.3) have been successfully adjusted. (Ⅳ) The influence of delay on the stability and the emergence of bifurcation of system (1.3) is explored. (Ⅴ) The control tactics can help us explore the control issue of Hopf bifurcation of plenty of differential dynamical models in lots of fields.

    We plan this study as follows. We explore the existence and uniqueness, non-negativeness of the solution to system (1.3) in Segment 2. The stability behavior and the emergence of bifurcation of system (1.3) are discussed in Segment 3. The control of Hopf bifurcation of system (1.3) via a proper hybrid controller including state feedback and parameter perturbation is dealt with in Segment 4. Segment 5 carries out computer simulations to support the correctness of the acquired outcomes. Segment 6 draws a simple conclusion.

    In this part, we consider the existence and uniqueness, boundedness of the solution to system (1.3) by making use of the fixed point theorem and inequality techniques.

    Theorem 2.1. Let Ω={u1,u2,u3,u4)R4:max{|u1|,|u2|,|u3|,|u4|}U}, where U represents a positive constant. For every (u10,u20,u30,u40)Ω, system (1.3) under the initial value (u10,u20,u30,u40) owns a unique solution U=(u1,u2,u3,u4)Ω.

    Proof. Let

    h(U)=(h1(U),h2(U),h3(U),h4(U)), (2.1)

    where

    {h1(U)=tanh[u4(t)u2(t)]au1(tθ),h2(U)=tanh[u1(t)+u4(t)]au2(tθ),h3(U)=tanh[u1(t)+u2(t)u4(t)]au3(tθ),h4(U)=tanh[u3(t)u2(t)]au4(tθ). (2.2)

    For every U,˜UΩ, one gains

    ||h(U)h(˜U)||=|tanh[u4(t)u2(t)]au1(tθ)[tanh[˜u4(t)˜u2(t)]a˜u1(tθ)]|+|tanh[u1(t)+u4(t)]au2(tθ)[tanh(˜u1(t)+˜u4(t))a˜u2(tθ)]|+|tanh[u1(t)+u2(t)u4(t)]au3(tθ)[tanh(˜u1(t)+˜u2(t)˜u4(t))a˜u3(tθ)]|+|tanh[u3(t)u2(t)]au4(tθ)[tanh(˜u3(t)˜u2(t))a˜u4(tθ))]|(2+a)|u1(t)˜u1(t)|+(3+a)|u2(t)˜u2(t)|+(1+a)|u3(t)˜u3(t)|+(3+a)|u4(t)˜u4(t)|(3+a)[|u1(t)˜u1(t)|+|u2(t)˜u2(t)|+|u3(t)˜u3(t)|+|u4(t)˜u4(t)|]=(3+a)||U˜U||. (2.3)

    Thus, h(U) conforms to Lipschitz condition with respect to U (see [17]). According to fixed point theorem, we can lightly conclude that Theorem 2.1 is right.

    Theorem 2.2. If θ=0, then every solution to system (1.3) starting with R4+ is uniformly bounded.

    Label

    U(t)=u1(t)+u2(t)+u3(t)+u4(t). (2.4)

    Then

    ˙U1(t)=˙u1(t)+˙u2(t)+˙u3(t)+˙u4(t)=tanh[u4(t)u2(t)]au1(t)+tanh[u1(t)+u4(t)]au2(t)+tanh[u1(t)+u2(t)u4(t)]au3(t)+tanh[u3(t)u2(t)]au4(tθ)a(u1(t)+u2(t)+u3(t)+u4(t))+4=aU(t)+4, (2.5)

    which results in

    ˙U1(t)+aU(t)4. (2.6)

    Then, one acquires

    U(t)4a,ast. (2.7)

    The proof of Theorem 2.2 comes to an end.

    Clearly, system (1.3) admits a unique zero equilibrium point: E(0,0,0,0). The linear system of system (1.3) at E(0,0,0,0) can be expressed as

    {˙u1(t)=u4(t)u2(t)au1(tθ),˙u2(t)=u1(t)+u4(t)au2(tθ),˙u3(t)=u1(t)+u2(t)u4(t)au3(tθ),˙u4(t)=u3(t)u2(t)au4(tθ). (3.1)

    The characteristic equation of system (3.1) owns the following expression:

    det[λ+aeλθ1011λ+aeλθ0111λ+aeλθ1011λ+aeλθ]=0, (3.2)

    which generates

    λ4+3λ2λ+1+(a1λ3+a2λ+a3)eλθ+(a4λ2+a5)e2λθ+a6λe3λθ+a7e4λθ=0, (3.3)

    where

    {a1=4a,a2=6a,a3=a,a4=6a2,a5=3a2a6=4a3,a7=a4. (3.4)

    When θ=0, then Eq (3.3) becomes:

    λ4+b1λ3+b2λ2+b3λ+b4=0, (3.5)

    where

    {b1=a1,b2=3+a4,b3=a2+a61,b4=1+a3+a5+a7. (3.6)

    If

    (A1){Δ1=b1>0,Δ2=det[b11b3b2]>0,Δ3=[b110b3b2b10b4b3]>0,Δ4=b4Δ3>0

    is fulfilled, then the four roots λ1,λ2,λ3,λ4 of Eq (3.5) have negative real parts. Thus the zero equilibrium point E(0,0,0,0) of system (1.3) with θ=0 is locally asymptotically stable.

    By (3.3), we get

    (λ4+3λ2λ+1)e2λθ+(a1λ3+a2λ+a3)eλθ+(a4λ2+a5)+a6λeλθ+a7e2λθ=0. (3.7)

    Suppose that λ=iϕ is the root of Eq (3.7). Then Eq (3.7) becomes:

    [(iϕ)4+3(iϕ)2iϕ+1]e2iϕθ+[a1(iϕ)3+a2iϕ+a3]eiϕθ+a4(iϕ)2+a5)+a6(iϕ)eiϕθ+a7e2iϕθ=0, (3.8)

    which generates

    (ϕ43iϕ2iϕ+1)(cos2ϕθ+isin2ϕθ)+(a1iϕ3+a2iϕ+a3)(cosϕθ+isinϕθ)+(a4iϕ2+a5)+a6iϕ(cosϕθisinϕθ)+a7(cos2ϕθisin2ϕθ)=0. (3.9)

    It follows from (3.9) that

    {(ϕ4+1+a7)cos2ϕθ+(3ϕ2+ϕ)sin2ϕθ+a3cosϕθ+(a6ϕa2+a1ϕ3)sinϕθ+a5=0,(ϕ4+1a7)sin2ϕθ(3ϕ2+ϕ)cos2ϕθ+(a2a1ϕ3)cosϕθ+(a3a6ϕ)sinϕθa4ϕ2=0. (3.10)

    Notice that sinϕθ=±1cos2ϕθ, then it follows from the first equation of system (3.10) that

    (ϕ4+1+a7)(2cos2ϕθ1)+(3ϕ2+ϕ)cosϕθ(±1cos2ϕθ)+a3cosϕθ+(a6ϕa2+a1ϕ3)(±1cos2ϕθ)+a5=0. (3.11)

    By (3.11), one derives

    [(ϕ4+1+a7)(2cos2ϕθ1)+a3cosϕθ+a5]2=(3ϕ2+ϕ)cosϕθ+(a6ϕa2+a1ϕ3)]2(1cos2ϕθ), (3.12)

    which leads to

    ϵ1cos4ϕθ+ϵ2cos3ϕθ+ϵ3cos2ϕθ+ϵ4cosϕθ+ϵ5=0, (3.13)

    where

    {ϵ1=4(ϕ4+1+a7)2+(3ϕ2+ϕ)2,ϵ2=2a3(ϕ4+1+a7)+2(3ϕ2+ϕ)(a6ϕa2+a1ϕ3),ϵ3=a23(3ϕ2+ϕ)2+(a6ϕa2+a1ϕ3)2,,ϵ4=2(a5ϕ41a7)2(3ϕ2+ϕ)(a6ϕa2+a1ϕ3),ϵ5=(a5ϕ41a7)2(a6ϕa2+a1ϕ3)2. (3.14)

    By (3.13), we can get the expression of cosϕθ. Here we assume that

    cosϕθ=f1(ϕ), (3.15)

    where f1(ϕ) is a function with respect to ϕ. In a similar way, we can also get the expression of sinϕθ. Here we assume that

    sinϕθ=f2(ϕ), (3.16)

    where f2(ϕ) is a function with respect to ϕ. Using (3.15) and (3.16), we get

    f21(ϕ)+f22(ϕ)=1. (3.17)

    We can obtain the root of (3.17) by computer. Here we denote the root of (3.1) by ϕ0. Using (3.15), one gains

    θ(h)0=1ϕ0[arccosf1(ϕ0)+2hπ], (3.18)

    where h=0,1,2,. Denote θ0=min{h=0,1,2,}{θ(h)0} and assume that when θ=θ0, (3.3) admits a pair of imaginary roots ±iϕ0.

    Now the following assumption is needed:

    (A2)U1RU2R+U1IU2I>0,

    where

    {U1R=1+(3a1ϕ20+a2)cosϕ0θ0+2a4ϕ0sin2ϕ0θ0+a6cos3ϕ0θ0,U1I=4ϕ306ϕ0(3a1ϕ20+a2)sinϕ0θ0+2a4ϕ0cos2ϕ0θ0a6sin3ϕ0θ0,U2R=a3ϕ0sinϕ0θ0(a1ϕ30a2ϕ0)ϕ0cosϕ0θ0+ϕ0(a5a4ϕ20)cos2ϕ0θ03a6ϕ20cos3ϕ0θ0+4a7ϕ0sin4ϕ0θ0,U2I=(a1ϕ30+a2ϕ0)ϕ0sinϕ0θ0+a2ϕ0cosϕ0θ0ϕ0(a5a4ϕ20)sin2ϕ0θ0+3a6ϕ20sin3ϕ0θ0+4a7ϕ0cos4ϕ0θ0. (3.19)

    Lemma 3.1. Let λ(θ)=ς1(θ)+iς2(θ) be the root of Eq (3.3) near θ=θ0 such that ς1(θ0)=0,ς2(θ0)=ϕ0, then Re(dλdθ)|θ=θ0,ϕ=ϕ0>0.

    Proof. In view of Eq (3.3), one gains

    (4λ36λ1)dλdθ+(3a1λ2+a2)eλθdλdθeλθ(dλdθθ+λ)(a1λ3+a2λ+a3)+2a4λdλdθe2λθe2λθ(dλdθθ+λ)(a4λ2+a5)+a6e3λθdλdθ3a6λe3λθ(dλdθθ+λ)4a7e4λθ(dλdθθ+λ)=0, (3.20)

    which leads to

    (dλdθ)1=U1(λ)U2(λ)θλ, (3.21)

    where

    {U1(λ)=4λ36λ+1+(3a1λ2+a2)eλθ+2a4λe2λθ+a6e3λθ,U2(λ)=λeλθ(a1λ3+a2λ+a3)+2λ(a4λ2+a5)e2λθ+3a6λ2e3λθ+4a7λe4λθ. (3.22)

    Then

    Re[(dλdθ)1]θ=θ0,ϕ=ϕ0=Re[U1(λ)U2(λ)]θ=θ0,ϕ=ϕ0=U1RU2R+U1IU2IU22R+U22I. (3.23)

    Taking advantage of (A2), one gets

    Re[(dλdθ)1]θ=θ0,ϕ=ϕ0>0. (3.24)

    The proof ends.

    Relying on the analysis above, we can lightly set up the following assertion.

    Theorem 3.1. If (A1),(A2) hold, then the zero equilibrium point E(0,0,0,0) of model (1.3) is locally asymptotically stable if θ[0,θ0) and model (1.3) produces a Hopf bifurcation around the zero equilibrium point E(0,0,0,0) when θ=θ0.

    Remark 3.1. In many cases, the self-feedback time delay of every neuron in neural networks is different. If there exist two different delays in model (1.3), we can also explore the Hopf bifurcation of system (1.3). We leave it for future work.

    In this part, we will deal with the Hopf bifurcation problem of system (1.5) via a suitable hybrid controller consisting of state feedback and parameter perturbation. Taking advantage of the idea in [19,20], we obtain the following controlled finance model:

    {˙u1(t)=tanh[u4(t)u2(t)]au1(tθ),˙u2(t)=tanh[u1(t)+u4(t)]au2(tθ),˙u3(t)=ρ1{tanh[u1(t)+u2(t)u4(t)]au3(tθ)}+ρ2u3(tθ),˙u4(t)=tanh[u3(t)u2(t)]au4(tθ), (4.1)

    where ρ1,ρ2 represent feedback gain parameters. Systems (4.1) and (1.3) have the same unique equilibrium point E(0,0,0,0). The linear system of system (4.1) at E(0,0,0,0) is given by

    {˙u1(t)=u4(t)u2(t)au1(tθ),˙u2(t)=u1(t)+u4(t)au2(tθ),˙u3(t)=ρ1u1(t)+ρ1u2(t)ρ1u4(t)(ρ1aρ2)u3(tθ),˙u4(t)=u3(t)u2(t)au4(tθ). (4.2)

    The characteristic equation of system (3.1) owns the following expression:

    det[λ+aeλθ1011λ+aeλθ01ρ1ρ1λ+(ρ1aρ2)eλθρ1011λ+aeλθ]=0, (4.3)

    which generates

    λ4+α1λ2+α2+(α3λ3+α4λ)eλθ+(α5λ2+α6)e2λθ+α7λe3λθ+α8e4λθ=0, (4.4)

    where

    {α1=2+ρ1,α2=ρ1,α3=3a+ρ1a+ρ2,α4=2(2a+ρ1a+ρ2),α5=3a(a+aρ1+ρ2),α6=a(2aρ1+2ρ2+a),α7=a2(3aρ1+3ρ2+a),α8=a3(aρ1+ρ2). (4.5)

    When θ=0, then Eq (4.4) becomes:

    λ4+β1λ3+β2λ2+β3λ+β4=0, (4.6)

    where

    {β1=α1,β2=α1+α5,β3=α4+α7,β4=α2+α6+α8. (4.7)

    If

    (A3){Λ1=β1>0,Λ2=det[β11β3β2]>0,Λ3=[β110β3β2β10β4β3]>0,Λ4=β4Λ3>0

    is fulfilled, then the four roots λ1,λ2,λ3,λ4 of Eq (4.4) have negative real parts. Thus the zero equilibrium point E(0,0,0,0) of system (4.1) with θ=0 is locally asymptotically stable.

    By (4.4), we get

    (λ4+α1λ2+α2)e2λθ+(α3λ3+α4λ)eλθ+(α5λ2+α6)+α7λeλθ+α8e2λθ=0. (4.8)

    Suppose that λ=iφ is the root of Eq (4.8). Then Eq (4.8) becomes:

    [(iφ)4+α1(iφ)2+α2]e2iφθ+[α3(iφ)3+α4(iφ)]eiφθ+[α5(iφ)2+α6]+α7iφeiφθ+α8e2iφθ=0, (4.9)

    which generates

    (φ4α1φ2+α2)(cos2φθ+isin2φθ)+(iα3φ3+iα4φ)(cosφθ+isinφθ)+(α5φ2+α6)+iα7φ(cosφθisinφθ)+α8(cos2φθisin2φθ)=0. (4.10)

    It follows from (4.10) that

    {(φ4α1φ2+α2+α8)cos2φθ+[(α7α4)φ+α3φ3]sinϕθ+α6α5φ2=0,(φ4α1φ2+α2α8)sin2φθ+(α4φα3φ3)cosφθ+α7φ=0. (4.11)

    In view of sinφθ=±1cos2φθ, then it follows from the second equation of (4.11) that

    2(φ4α1φ2+α2α8)(±1cos2φθ)cosφθ+(α4φα3φ3)cosφθ+α7φ=0, (4.12)

    which leads to

    ϵ1cos4φθ+ϵ2cos2φθ+ϵ3cosφθ+ϵ4=0, (4.13)

    where

    {ϵ1=4(φ4α1φ+α2α8)2,ϵ2=(α4φα3φ3)24(φ4α1φ+α2α8)2,ϵ3=2α7φ(α4φα3φ3),ϵ4=α27φ2. (4.14)

    By virtue of computer software, we can solve the value of cosφθ. Here we assume that

    cosφθ=h1(φ). (4.15)

    Then we can solve the value of sinφθ. Here we assume that

    sinφθ=h2(φ). (4.16)

    Using (4.15) and (4.16), we have

    h21(φ)+h22(φ)=1. (4.17)

    By virtue of (4.17), we can easily obtain the value of φ (say φ0) via computer software. According to (4.15), we get

    θj=1φ0[arccosf1(φ0)+2jπ],j=0,1,2,. (4.18)

    Denote

    θ=min{j=0,1,2,}{θj}. (4.19)

    Now we know that when θ=θ, (4.4) has a pair of imaginary roots ±iφ0.

    Now we give the following hypothesis:

    (A4)V1RV2R+V1IV2I>0,

    where

    {V1R=(α42α3φ20)cosφ0θ+2α5φ0sinφ0θ,V1I=2α1ς04φ30(α42α3φ20)sinφ0θ+2α5φ0cosφ0θ,V2R=(α3φ30α4φ0)φ0cosφ0θ+2φ0(α6α5φ20)sin2φ0θ3α7φ20cos3φ0θ+4α8φ0sin4φ0θ,V2I=(α4φ0α3φ30)φ0sinφ0θ+2φ0(α6α5φ20)cos2φ0θ3α7φ20sin3φ0θ+4α8φ0cos4φ0θ. (4.20)

    Lemma 4.1. Let λ(θ)=ξ1(θ)+iξ2(θ) be the root of Eq (4.4) around θ=θ satisfying ξ1(θ)=0,ξ2(θ)=φ0, then Re(dλdθ)|θ=θ,φ=φ0>0.

    Proof. By Eq (4.4), one gets

    (3λ2+2α1λ)dλdθ+(2α3λ2+α4)eλθdλdθeλθ(dλdθθ+λ)(α3λ3+α4λ)+2α5λe2λθdλdθ2e2λθ(dλdθθ+λ)(α5λ2+α6)+α7e3λθ(dλdθθ+λ)3e3λθ(dλdθθ+λ)α7λ4α8e4λθ(dλdθθ+λ)=0, (4.21)

    which implies

    (dλdσ)1=V1(λ)V2(λ)θλ, (4.22)

    where

    {V1(λ)=4λ3+2α1λ+(2α3λ2+α4)eλθ+2α5λe2λσ,V2(λ)=λeλθ(α3λ3+α4λ)+2λe2λθ(α5λ2+α6)+3λ2α7e3λσ+4α8λe4λσ. (4.23)

    Hence

    Re[(dλdθ)1]θ=θ,φ=φ0=Re[V1(λ)V2(λ)]θ=θ,φ=φ0=V1RV2R+V1IV2IV22R+V22I. (4.24)

    Taking advantage of (A4), one gets

    Re[(dλdσ)1]θ=θ,φ=φ0>0, (4.25)

    which ends the proof.

    Relaying on the discussion above, we can lightly gain the following outcome.

    Theorem 4.1. If (A3),(A4) hold, then the equilibrium point E(0,0,0,0) of model (4.1) is locally asymptotically stable if θ[0,θ) and a Hopf bifurcation of model (4.1) arises around the equilibrium point E(0,0,0,0) when θ=θ.

    Remark 4.1. In 2008, Sprott [13] investigated the chaotic behavior of the 4D neural network model (1.1); In 2022, Vaishwar and Yadav [14] dealt with the stability and Hopf-bifurcation of the 4D neural network model (1.2). Both works of [13] and [14] are only concerned with the integer-order neural networks. In this study, we deal with the stability, Hopf bifufcation and its control issue of the 4D fractional-order neural network model (1.3) concerning delay. All the exploration ways are different from those in [13] and [14]. Relying on this viewpoint, we argue that the exploration ways and the acquired outcomes are important supplement to the works of [13] and [14].

    Remark 4.2. In this paper, we have dealt with the Hopf bifurcation control via hybrid controller. Of course, we can also deal with the Hopf bifurcation control via other hybrid controllers. We will explore this topic in near future.

    Example 5.1. Consider the following 4D BAM neural networks concerning time delay:

    {˙u1(t)=tanh[u4(t)u2(t)]0.25u1(tθ),˙u2(t)=tanh[u1(t)+u4(t)]0.25u2(tθ),˙u3(t)=tanh[u1(t)+u2(t)u4(t)]0.25u3(tθ),˙u4(t)=tanh[u3(t)u2(t)]0.25u4(tθ). (5.1)

    It is not difficult to gain that system (5.1) has a unique zero equilibrium point E(0,0,0,0). One is lightly able to verify that both conditions (A1) and (A2) of Theorem 3.1 hold true. Making use of computer software, one gains ϕ0=2.0083,θ00.8. To examine the accuracy of the gained outcomes of Theorem 3.1, we give both different time delay values. Let θ=0.76 and θ=1.1. For θ=0.76<θ00.8, the computer simulation outcomes are provided in Figure 1. In terms of Figure 1, one can easily understand that u10,u20,u30,u40 when t+. In other words, the zero equilibrium point E(0,0,0,0) of system (5.1) maintains locally asymptotically stable situation. For θ=1.1>θ00.8, the computer simulation outcomes are provided in Figure 2. In terms of Figure 2, one can easily understand that u1,u2,u3,u4 are going to maintain periodic motion around the value 0, respectively. In other words, a cluster of limit cycles (i.e., Hopf bifurcations) happen around the zero equilibrium point E(0,0,0,0). In addition, the bifurcation plots, which clearly display the bifurcation value of system (5.1), are provided in Figures 36. In terms of Figures 36, one is lightly able to understand that the bifurcation point of system (5.1) is θ00.8.

    Figure 1.  Computer simulation figures of system (5.1) concerning the delay θ=0.76<θ00.8. The zero equilibrium point E(0,0,0,0) maintains locally asymptotically stable situation.
    Figure 2.  Computer simulation figures of system (5.1) concerning the delay θ=1.1>θ00.8. A cluster of limit cycles (i.e., Hopf bifurcations) happen near the zero equilibrium point E(0,0,0,0).
    Figure 3.  Bifurcation figure of system (5.1): x-axis denotes the time t and y-axis denotes the state variable u1. The bifurcation point θ00.8.
    Figure 4.  Bifurcation figure of system (5.1): x-axis denotes the time t and y-axis denotes the state variable u2. The bifurcation point θ00.8.
    Figure 5.  Bifurcation figure of system (5.1): x-axis denotes the time t and y-axis denotes the state variable u3. The bifurcation point θ00.8.
    Figure 6.  Bifurcation figure of system (5.1): x-axis denotes the time t and y-axis denotes the state variable u4. The bifurcation point θ00.8.

    Example 5.2. Consider the following controlled 4D BAM neural networks concerning time delay:

    {˙u1(t)=tanh[u4(t)u2(t)]0.25u1(tθ),˙u2(t)=tanh[u1(t)+u4(t)]0.25u2(tθ),˙u3(t)=ρ1{tanh[u1(t)+u2(t)u4(t)]0.25u3(tθ)}+ρ2u3(tθ),˙u4(t)=tanh[u3(t)u2(t)]0.25u4(tθ). (5.2)

    It is not difficult to gain that system (5.2) has a unique zero equilibrium point E(0,0,0,0). One is lightly able to verify that both conditions (A3) and (A4) of Theorem 4.1 hold true. Making use of computer software, one gains φ0=5.6712,θ0.62. To examine the accuracy of the gained outcomes of Theorem 4.1, we give both different time delay values. Let θ=0.6 and θ=0.7. For θ=0.6<θ00.62, the computer simulation outcomes are provided in Figure 7. In terms of Figure 7, one can easily understand that u10,u20,u30,u40 when t+. In other words, the zero equilibrium point E(0,0,0,0) of system (5.2) maintains locally asymptotically stable situation. For θ=0.7>θ0.62, the computer simulation outcomes are provided in Figure 8. In terms of Figure 8, one can easily understand that u1,u2,u3,u4 are going to maintain periodic motion around the value 0, respectively. In other words, a cluster of limit cycles (i.e., Hopf bifurcations) happen around the zero equilibrium point E(0,0,0,0). In addition, the bifurcation plots, which clearly display the bifurcation value of system (5.2), are provided in Figures 912. In terms of Figures 912, one is lightly able to understand that the bifurcation point of system (5.2) is θ0.62.

    Figure 7.  Computer simulation figures of system (5.2) concerning the delay θ=0.6<θ0.62. The zero equilibrium point E(0,0,0,0) maintains locally asymptotically stable situation.
    Figure 8.  Computer simulation figures of system (5.2) concerning the delay θ=0.7>θ0.62. A cluster of limit cycles (i.e., Hopf bifurcations) happen near the zero equilibrium point E(0,0,0,0).
    Figure 9.  Bifurcation figure of system (5.2): x-axis denotes the time t and y-axis denotes the state variable u1. The bifurcation point θ0.62.
    Figure 10.  Bifurcation figure of system (5.2): x-axis denotes the time t and y-axis denotes the state variable u2. The bifurcation point θ0.62.
    Figure 11.  Bifurcation figure of system (5.2): x-axis denotes the time t and y-axis denotes the state variable u3. The bifurcation point θ0.62.
    Figure 12.  Bifurcation figure of system (5.2): x-axis denotes the time t and y-axis denotes the state variable u4. The bifurcation point θ0.62.

    Remark 5.1. From the computer simulation outcomes of Examples 5.1 and 5.2, we know that the bifurcation value of system (5.1) is θ00.8 and the bifurcation value of system (5.2) is θ0.62. Thus we understand that the stability domain of system (5.1) is narrowed and the time of onset of Hopf bifurcation of system (5.1) is advanced via a suitable hybrid controller consisting of state feedback and parameter perturbation.

    Delayed-driven Hopf bifurcation is a vital dynamical behavior in delayed dynamical models. In particular, delayed-driven Hopf bifurcation in neural networks has attracted great interest from scientific community. In this study, novel 4D BAM neural networks involving time delay have been formulated. The existence and uniqueness, boundedness of solution of the addressed BAM neural networks concerning single delay have been explored. Regarding the delay as Hopf bifurcation parameter, a new delay-independent condition on the stability and the emergence of delay-driven Hopf bifurcation is acquired. The importance of delay in stabilizing networks and dominating the time of onset of delay-driven Hopf bifurcation is reveled. In order to adjust the stability domain and the time of delay-driven Hopf bifurcation of addressed BAM neural networks concerning single delay, we apply a proper hybrid controller including state feedback and parameter perturbation to explore the stability domain and the time of delay-driven Hopf bifurcation issue. The study shows that the designed hybrid controller is an effective controller. The acquired outcomes of this study have momentous theoretical significance in controlling and contriving neural networks. Also the obtained results have important application in artificial intelligence in various industrial applications, especially in terms of convolutional fusion framework for collaborative fault identification of rotating machinery, and Physics-Informed Residual Network (PIResNet) for rolling element bearing fault diagnostics [39]. What is more, the exploration idea of this study is able to investigate the Hopf bifurcation and its control issue in lots of differential dynamical systems. In this paper, we have dealt with the Hopf bifurcation and its control of integer-order neural networks via hybrid controller. Of course, we can also deal with the Hopf bifurcation and its control of fractional-order neural networks via hybrid controller [40,41,42,43,44,45,46,47]. We will explore this topic in near future.

    This work is supported by National Natural Science Foundation of China (No.12261015, No.62062018), Project of High-level Innovative Talents of Guizhou Province ([2016]5651).

    The authors would like to thank the referees and the editor for helpful suggestions incorporated into this paper.

    The authors declare that they have no conflict of interest.



    [1] R. Zhao, B. X. Wang, J. G. Jian, Global μ-stabilization of quaternion-valued inertial BAM neural networks with time-varying delays via time-delayed impulsive control, Math. Comput. Simul., 202 (2022), 223–245. http://dx.doi.org/10.1016/j.matcom.2022.05.036 doi: 10.1016/j.matcom.2022.05.036
    [2] C. J. Xu, D. Mu, Y. L. Pan, C. Aouiti, Y. C. Pang, L. Y. Yao, Probing into bifurcation for fractional-order BAM neural networks concerning multiple time delays, J. Computat. Sci., 62 (2022), 101701. http://dx.doi.org/10.1016/j.jocs.2022.101701 doi: 10.1016/j.jocs.2022.101701
    [3] J. J. Oliveira, Global stability criteria for nonlinear differential systems with infinite delay and applications to BAM neural networks, Chaos Soliton. Fract., 164 (2022), 112676. http://dx.doi.org/10.1016/j.chaos.2022.112676 doi: 10.1016/j.chaos.2022.112676
    [4] D. Z. Chen, Z. Q. 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
    [5] X. Y. Mao, X. M. Wang, H. Y. Qin, Stability analysis of quaternion-valued BAM neural networks fractional-order model with impulses and proportional delays, Neurocomputing, 509 (2022), 206–220. http://dx.doi.org/10.1016/j.neucom.2022.08.059 doi: 10.1016/j.neucom.2022.08.059
    [6] C. J. Xu, M. Farman, Z. X. Liu, Y. C. Pang, Numerical approximation and analysis of epidemic model with constant proportional caputo(CPC) operator, Fractals, 32 (2024), 2440014. http://dx.doi.org/10.1142/S0218348X24400140 doi: 10.1142/S0218348X24400140
    [7] P. L. Li, C. J. Xu, M. Farman, A. Akgul, Y. C. Pang, Qualitative and stability analysis of fractional order emotion panic spreading model insight of fractional operator, Fractals, 32 (2024), 2440011. http://dx.doi.org/10.1142/S0218348X24400115 doi: 10.1142/S0218348X24400115
    [8] 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. Control, 27 (2022), 1030–1053. http://dx.doi.org/10.15388/namc.2022.27.28491 doi: 10.15388/namc.2022.27.28491
    [9] Y. Cao, S. Ramajayam, R. Sriraman, R. Samidurai, Leakage delay on stabilization of finite-time complex-valued BAM neural network: Decomposition approach, Neurocomputing, 463 (2021), 505–513. http://dx.doi.org/10.1016/j.neucom.2021.08.056 doi: 10.1016/j.neucom.2021.08.056
    [10] M. Syed Ali, G. Narayanan, V. Shekher, A. Alsaedi, B. Ahmad, Global Mittag-Leffler stability analysis of impulsive fractional-order complex-valued BAM neural networks with time varying delays, Commun. Nonlinear Sci. Numer. Simul., 83 (2020), 105088. http://dx.doi.org/10.1016/j.cnsns.2019.105088 doi: 10.1016/j.cnsns.2019.105088
    [11] B. Kosko, Bidirectional associative memories, IEEE Trans. Syst. Man Cyber., 18 (1988), 49–60. http://dx.doi.org/10.1109/21.87054 doi: 10.1109/21.87054
    [12] W. G. Yang, Periodic solution for fuzzy Cohen-Grossberg bam neural networks with both time-varying and distributed delays and variable coefficients, Neural Process. Lett., 40 (2014), 51–73. http://dx.doi.org/10.1007/s11063-013-9310-0 doi: 10.1007/s11063-013-9310-0
    [13] J. Sprott, Chaotic dynamics on large networks, Chaos, 18 (2008), 023135. http://dx.doi.org/10.1063/1.2945229 doi: 10.1063/1.2945229
    [14] A. Vaishwar, B. K. Yadav, Stability and Hopf-bifurcation analysis of four dimensional minimal neural network model with multiple time delays, Chinese J. Phys., 77 (2022), 300–318. http://dx.doi.org/10.1016/j.cjph.2022.02.011 doi: 10.1016/j.cjph.2022.02.011
    [15] B. P. Belousov, A periodic reaction and its mechanism, New York: John Wiley, 1985.
    [16] A. M. Zhabotinskii, Periodic process of the oxidation of malonic acid in solution (Study of the kinetics of Belousov,s), Biofizika, 9 (1964), 306–311.
    [17] Q. Din, T. Donchev, D. Kolev, Stability, Bifurcation analysis and chaos control in chlorine dioxide-iodine-malonic acid reaction, MATCH Commun. Math. Comput. Chem., 79 (2018), 577–606.
    [18] I. Lengyel, G. Ribai, I. R. Epstein, Experimental and modeling study of oscillations in the chlorine dioxide-iodine-malonic acid reaction, J. Amer. Chem. Soc., 112 (1990), 9104–9110. http://dx.doi.org/10.1021/ja00181a011 doi: 10.1021/ja00181a011
    [19] E. Mosekilde, Topics in Nonlinear Dynamics: Applications to Physics, Biology and Economic Systems, New Jersey: World Science Publisher, 1996. http://dx.doi.org/10.1142/3194
    [20] J. N. Wang, H. B. Shi, L. Xu, L. Zang, Hopf bifurcation and chaos of tumor-Lymphatic model with two time delays, Chaos Soliton. Fract., 157 (2022), 111922. http://dx.doi.org/10.1016/j.chaos.2022.111922 doi: 10.1016/j.chaos.2022.111922
    [21] N. C. Pati, B. Ghosh, Delayed carrying capacity induced subcritical and supercritical Hopf bifurcations in a predator-prey system, Math. Comput. Simul., 195 (2022), 171–196. http://dx.doi.org/10.1016/j.matcom.2022.01.008 doi: 10.1016/j.matcom.2022.01.008
    [22] C. J. Xu, Z. X. Liu, M. X. Liao, L. Y. Yao, Theoretical analysis and computer simulations of a fractional order bank data model incorporating two unequal time delays, Expert Syst. Appl., 199 (2022), 116859. http://dx.doi.org/10.1016/j.eswa.2022.116859 doi: 10.1016/j.eswa.2022.116859
    [23] C. J. Xu, W. Zhang, C. Aouiti, Z. X. Liu, L. Y. Yao, Further analysis on dynamical properties of fractional-order bi-directional associative memory neural networks involving double delays, Math. Meth. Appl. Sci., 45 (2022), 11736–11754. http://dx.doi.org/10.1002/mma.8477 doi: 10.1002/mma.8477
    [24] 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://dx.doi.org/10.1007/s12559-020-09782-w doi: 10.1007/s12559-020-09782-w
    [25] C. J. Xu, W. Zhang, Z. X. Liu, L. Y. Yao, Delay-induced periodic oscillation for fractional-order neural networks with mixed delays, Neurocomputing, 488 (2022), 681–693. http://dx.doi.org/10.1016/j.neucom.2021.11.079 doi: 10.1016/j.neucom.2021.11.079
    [26] C. J. Xu, D. Mu, Z. X. Liu, Y. C. Pang, M. X. Liao, P. Li, Bifurcation dynamics and control mechanism of a fractional-order delayed Brusselator chemical reaction model, MATCH Commun. Math. Comput. Chem., 89 (2023), 73–106. http://dx.doi.org/10.46793/match.89-1.073X doi: 10.46793/match.89-1.073X
    [27] C. J. Xu, C. Aouiti, Z. X. Liu, P. L. Li, L. Y. Yao, Bifurcation caused by delay in a fractional-order coupled Oregonator model in chemistry, MATCH Commun. Math. Comput. Chem., 88 (2022), 371–396. http://dx.doi.org/10.46793/match.88-2.371X doi: 10.46793/match.88-2.371X
    [28] C. J. Xu, W. Zhang, C. Aouiti, Z. X. Liu, P. L. Li, Bifurcation dynamics in a fractional-order Oregonator model including time delay, MATCH Commun. Math. Comput. Chem., 87 (2022), 397–414. http://dx.doi.org/10.46793/match.87-2.397X doi: 10.46793/match.87-2.397X
    [29] C. J. Xu, J. T. Lin, Y. Y. Zhao, Q. Y. Cui, W. Ou, Y. C. Pang, et al., New results on bifurcation for fractional-order octonion-valued neural networks involving delays, Netw. Comput. Neural Syst., 2024. https://doi.org/10.1080/0954898X.2024.2332662 doi: 10.1080/0954898X.2024.2332662
    [30] C. R. Tian, Y. Liu, Delay-driven Hopf bifurcation in a networked Malaria model, Appl. Mathe. Lett., 132 (2022), 108092. http://dx.doi.org/10.1016/j.aml.2022.108092 doi: 10.1016/j.aml.2022.108092
    [31] H. L. Li, L. Zhang, C. Hu, Y. L. Jiang, Z. D. Teng, Dynamical analysis of a fractional-order prey-predator model incorporating a prey refuge, J. Appl. Math. Comput., 54 (2017), 435–449. http://dx.doi.org/10.1007/s12190-016-1017-8 doi: 10.1007/s12190-016-1017-8
    [32] M. Das, A. Maiti, G. P. Samanta, Stability analysis of a prey-predator fractional order model incorporating prey refuge, Ecol. Genet. Genom., 7–8 (2018), 33–46. http://dx.doi.org/10.1016/j.egg.2018.05.001 doi: 10.1016/j.egg.2018.05.001
    [33] Z. Z. Zhang, H. Z. Yang, Hybrid control of Hopf bifurcation in a two prey one predator system with time delay, Proc. Chinese Contr. Conf., 2014, 6895–6900. http://dx.doi.org/10.1109/ChiCC.2014.6896136 doi: 10.1109/ChiCC.2014.6896136
    [34] L. P. Zhang, H. N. Wang, M. Xu, Hybrid control of bifurcation in a predator-prey system with three delays, Acta Phys. Sinica, 60 (2011), 010506. http://dx.doi.org/10.7498/aps.60.010506 doi: 10.7498/aps.60.010506
    [35] Z. Liu, K. W. Chuang, Hybrid control of bifurcation in continuous nonlinear dynamical systems, Int. J. Bifur. Chaos, 15 (2005), 1895–3903. http://dx.doi.org/10.1142/S0218127405014374 doi: 10.1142/S0218127405014374
    [36] Y. Y. Ni, Z. Wang, X. Huang, Q. Ma, H. Shen, Intermittent sampled-data control for local stabilization of neural networks subject to actuator saturation: A work-interval-dependent functional approach, IEEE Trans. Neural Netw. Learn. Syst., 35 (2024), 1087–1097. http://dx.doi.org/10.1109/TNNLS.2022.3180076 doi: 10.1109/TNNLS.2022.3180076
    [37] L. Yao, Z. Wang, X. Huang, Y. X. Li, Q. Ma, H. Shen, Stochastic sampled-data exponential synchronization of Markovian jump neural networks with time-varying delay, IEEE Trans. Neural Netw. Learn. Syst., 34 (2023), 909–920. http://dx.doi.org/10.1109/TNNLS.2021.3103958 doi: 10.1109/TNNLS.2021.3103958
    [38] X. D. Si, Z. Wang, Y. J. Fan, H. Shen, Sampled-data-based bipartite leader-following synchronization of cooperation-competition neural networks via scheduled-interval looped-function, IEEE Trans. Circuits Syst. I, 70 (2023), 3723–3734. http://dx.doi.org/10.1109/TCSI.2023.3284858 doi: 10.1109/TCSI.2023.3284858
    [39] Q. Ni, J. C. Ji, B. Halkon, K. Feng, A. K. Nandi, Physics-Informed Residual Network (PIResNet) for rolling element bearing fault diagnostics, Mech. Syst. Signal Process., 200 (2023), 110544. http://dx.doi.org/10.1016/j.ymssp.2023.110544 doi: 10.1016/j.ymssp.2023.110544
    [40] P. L. Li, R. Gao, C. J. Xu, J. W. Shen, S. Ahmad, Y. Li, Exploring the impact of delay on Hopf bifurcation of a type of BAM neural network models concerning three nonidentical delays, Neural Process. Lett., 55 (2023), 5905–5921. http://dx.doi.org/10.1007/s11063-023-11392-0 doi: 10.1007/s11063-023-11392-0
    [41] C. J. Xu, Y. Y. Zhao, J. T. Lin, Y. C. Pang, Z. X. Liu, J. W. Shen, et al., Mathematical exploration on control of bifurcation for a plankton-oxygen dynamical model owning delay, J. Math. Chem., 2023, http://dx.doi.org/10.1007/s10910-023-01543-y doi: 10.1007/s10910-023-01543-y
    [42] W. Ou, C. J. Xu, Q. Y. Cui, Y. C. Pang, Z. X. Liu, J. W. Shen, et al., Hopf bifurcation exploration and control technique in a predator-prey system incorporating delay, AIMS Math., 9 (2023), 1622–1651. http://dx.doi.org/10.3934/math.2024080 doi: 10.3934/math.2024080
    [43] Q. Y. Cui, C. J. Xu, W. Ou, Y. C. Pang, Z. X. Liu, P. L. Li, et al., Bifurcation behavior and hybrid controller design of a 2D Lotka-Volterra commensal symbiosis system accompanying delay, Mathematics, 11 (2023), 4808. http://dx.doi.org/10.3390/math11234808 doi: 10.3390/math11234808
    [44] C. Maharajan, C. Sowmiya, C. J. Xu, Fractional order uncertain BAM neural networks with mixed time delays: An existence and Quasi-uniform stability analysis, J. Intell. Fuzzy Syst., 46 (2024), 4291–4313. http://dx.doi.org/10.3233/JIFS-234744 doi: 10.3233/JIFS-234744
    [45] C. J. Xu, M. Farman, A. Shehzad, Analysis and chaotic behavior of a fish farming model with singular and non-singular kernel, Inter. J. Biomath., 2023. http://dx.doi.org/10.1142/S179352452350105X doi: 10.1142/S179352452350105X
    [46] C. J. Xu, Y. C. Pang, Z. X. Liu, J. W. Shen, M. X. Liao, P. L. Li, Insights into COVID-19 stochastic modelling with effects of various transmission rates: Simulations with real statistical data from UK, Australia, Spain, and India, Phys. Scripta, 99 (2024), 025218. http://dx.doi.org/10.1088/1402-4896/ad186c doi: 10.1088/1402-4896/ad186c
    [47] C. J. Xu, M. X. Liao, P. L. Li, L. Y. Yao, Q. W. Qin, Y. L. Shang, Chaos control for a fractional-order Jerk system via time delay feedback controller and mixed controller, Fract. Fractional, 5 (2021), 257. http://dx.doi.org/10.3390/fractalfract5040257 doi: 10.3390/fractalfract5040257
  • This article has been cited by:

    1. Qingyi Cui, Changjin Xu, Yiya Xu, Wei Ou, Yicheng Pang, Zixin Liu, Jianwei Shen, Muhammad Zafarullah Baber, Chinnamuniyandi Maharajan, Uttam Ghosh, Bifurcation and Controller Design of 5D BAM Neural Networks With Time Delay, 2024, 37, 0894-3370, 10.1002/jnm.3316
  • 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(1016) PDF downloads(42) Cited by(1)

Figures and Tables

Figures(12)

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog