Research article Special Issues

New criteria on global Mittag-Leffler synchronization for Caputo-type delayed Cohen-Grossberg Inertial Neural Networks

  • Our focus of this paper is on global Mittag-Leffler synchronization (GMLS) of the Caputo-type Inertial Cohen-Grossberg Neural Networks (ICGNNs) with discrete and distributed delays. This model takes into account the inertial term as well as the two types of delays, which greatly reduces the conservatism with respect to the model. A change of variables transforms the 2β order inertial frame into β order ordinary frame in order to deal with the effect of the inertial term. In the following steps, two novel types of delay controllers are designed for the purpose of reaching the GMLS. In conjunction with the novel controllers, utilizing differential mean-value theorem and inequality techniques, several criteria are derived to determine the GMLS of ICGNNs within the framework of Caputo-type derivative and calculus properties. At length, the feasibility of the results is further demonstrated by two simulation examples.

    Citation: Hongmei Zhang, Xiangnian Yin, Hai Zhang, Weiwei Zhang. New criteria on global Mittag-Leffler synchronization for Caputo-type delayed Cohen-Grossberg Inertial Neural Networks[J]. AIMS Mathematics, 2023, 8(12): 29239-29259. doi: 10.3934/math.20231497

    Related Papers:

    [1] Ruoyu Wei, Jinde Cao, Wenhua Qian, Changfeng Xue, Xiaoshuai Ding . Finite-time and fixed-time stabilization of inertial memristive Cohen-Grossberg neural networks via non-reduced order method. AIMS Mathematics, 2021, 6(7): 6915-6932. doi: 10.3934/math.2021405
    [2] Pratap Anbalagan, Evren Hincal, Raja Ramachandran, Dumitru Baleanu, Jinde Cao, Chuangxia Huang, Michal Niezabitowski . Delay-coupled fractional order complex Cohen-Grossberg neural networks under parameter uncertainty: Synchronization stability criteria. AIMS Mathematics, 2021, 6(3): 2844-2873. doi: 10.3934/math.2021172
    [3] Mohammed D. Kassim . A fractional Halanay inequality for neutral systems and its application to Cohen-Grossberg neural networks. AIMS Mathematics, 2025, 10(2): 2466-2491. doi: 10.3934/math.2025115
    [4] Jingfeng Wang, Chuanzhi Bai . Global Mittag-Leffler stability of Caputo fractional-order fuzzy inertial neural networks with delay. AIMS Mathematics, 2023, 8(10): 22538-22552. doi: 10.3934/math.20231148
    [5] Yijia Zhang, Tao Xie, Yunlong Ma . Robustness analysis of exponential stability of Cohen-Grossberg neural network with neutral terms. AIMS Mathematics, 2025, 10(3): 4938-4954. doi: 10.3934/math.2025226
    [6] Huizhen Qu, Jianwen Zhou . $ S $-asymptotically $ \omega $-periodic dynamics in a fractional-order dual inertial neural networks with time-varying lags. AIMS Mathematics, 2022, 7(2): 2782-2809. doi: 10.3934/math.2022154
    [7] 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
    [8] Tao Xie, Wenqing Zheng . Robustness analysis of Cohen-Grossberg neural network with piecewise constant argument and stochastic disturbances. AIMS Mathematics, 2024, 9(2): 3097-3125. doi: 10.3934/math.2024151
    [9] Yang Xu, Zhouping Yin, Yuanzhi Wang, Qi Liu, Anwarud Din . Mittag-Leffler projective synchronization of uncertain fractional-order fuzzy complex valued neural networks with distributed and time-varying delays. AIMS Mathematics, 2024, 9(9): 25577-25602. doi: 10.3934/math.20241249
    [10] Biwen Li, Yibo Sun . Stability analysis of Cohen-Grossberg neural networks with time-varying delay by flexible terminal interpolation method. AIMS Mathematics, 2023, 8(8): 17744-17764. doi: 10.3934/math.2023906
  • Our focus of this paper is on global Mittag-Leffler synchronization (GMLS) of the Caputo-type Inertial Cohen-Grossberg Neural Networks (ICGNNs) with discrete and distributed delays. This model takes into account the inertial term as well as the two types of delays, which greatly reduces the conservatism with respect to the model. A change of variables transforms the 2β order inertial frame into β order ordinary frame in order to deal with the effect of the inertial term. In the following steps, two novel types of delay controllers are designed for the purpose of reaching the GMLS. In conjunction with the novel controllers, utilizing differential mean-value theorem and inequality techniques, several criteria are derived to determine the GMLS of ICGNNs within the framework of Caputo-type derivative and calculus properties. At length, the feasibility of the results is further demonstrated by two simulation examples.



    Fractional calculus, as a prominent branch of mathematics, traces back to the 17th century [1]. It has been extensively used in environmental mechanics, engineering applications, automatic control and signal processing [2,3,4,5]. In contrast to integer order integrals, fractional order integrals can additionally describe procedures of all kinds with memory and hereditary properties. It was first developed in the 1940s to model neural computation, which is the foundation of Neural Networks (NNs) models. In the last few years, NNs have received significant attention from several scholars, and rich results are available in the reports [6,7,8,9,10].

    In 1986, as can be noticed, Babcock and Westervelt brought inductance into a class of circuit models, resulting in inertial NNs (INNs). As a particular class of NNs, INNs take the form of second-order differential systems. The inclusion of the inertial term gives rise to some complicated behavior, such as instabilities and chaos. Therefore, INNs have caught sufficient attention of numerous researchers to yield some fascinating results. It is not difficult to see from these studies that the treatment of the inertial term is in general a method of order reduction or direct analysis. [8] discussed the problem of stabilization of INNs, which used analysis approach. [9] considered the inertial system by the method of non-separation. Based on integer order INNs, various synchronization and stability problems were widely discussed [9,10,11,12,13,14]. Compared to integer-order NNs, fractional-order NNs (FONNs) have the property of infinite storage and have the superior ability to describe dynamic behavior. As a result, a growing number of scholars have introduced inertial terms into fractional systems to form FOINNs, with a plethora of intriguing results [15,16,17,18]. In actual applications, INNs are widely used in engineering, such as secure communication [19], image encryption [20], etc.

    It may be noted that the CGNNs were raised by Cohen and Grossberg as a generalization of the network models. This class of exceptional NNs model has attracted considerable attention in the areas of associative memory, classification, solving optimization problems and pattern recognition [21,22,23,24]. Recently, a large amount of results for CGNNs have been reported. In [25], the problem of synchronization of complex-valued CGNNs is addressed by constructing appropriate controllers. In [26], proposed the model of delayed fuzzy CGNNs with discontinuous activations and uncertain term. Due to the particularity of INNs and CGNNs, some scholars add inertial term to CGNNs to form a kind of inertial CGNNs (ICGNNs) and the dynamic behavior analysis of this class of NNs has received wide attention [8,27,28,29]. For example, [8] proposed the condition of judging the stability by the direct approach of ICGNNs with generalized delays. [27] gained several new criteria to verify the ICGNNs with proportional delays. [28] established the criterion of fixed-time synchronization for ICGNNs by constructing a controller with two exponential terms. To better describe the dynamic behavior of some neurons, [30] proposed the delayed of FOICGNNs. As a matter of fact, some results of FOCGNNs have been reported [31,32,33,34], where the models do not take into account the inertia term. Furthermore, there are some systems of FOINNs that are not combined with CGNNs [15,16,17,18]. It should be pointed out that model is based on the Riemann-Liouville type FOICGNNs in [30]. Thus, we found that there is relatively little research on FOICGNNs in the Caputo sense. It is noteworthy that various types of synchronization have been extensively discussed, including but not limited to anti-synchronization [12], fixed-time synchronization [7,11,19,26,28], finite time synchronization [25,35,36], quasi projective synchronization [25], global asymptotic synchronization (GAS) [16,27] and GMLS [17,18]. Currently, various effective control strategies have been applied to solve the synchronization problem, such as impulsive control [37], predefined-time control [12], linear control [25,29], pinning control [13,38] and sliding mode control [14]. Therefore, the problem of GMLS for Caputo-type FOICGNNs with discrete and distributed delays is not yet considered.

    Motivated by the front statement, our main goal of this paper is to study the GMLS of Caputo-type delayed ICGNNs. The following innovations are listed:

    (1) The model being considered is constructed on the basis of Caputo-type ICGNNs that have both discrete and distributed delays. The model takes into account two distinct time delays, thus making it more comprehensive and more practical than the models of ICGNNs [28,29].

    (2) The inertial system is converted into two ordinary fractional systems by variable substitution. The difficulties caused by delays are overcome using two novel control strategies. Different from [30], the novel controllers are designed to achieve the GMLS of Caputo-type ICGNNs, which reduces the control cost and improves the control precision to a certain extent.

    (3) Compared with the matrix inequalities approach [32], the novel conditions of GMLS are derived by the algebraic inequality technique, which can avoid the complex calculation and facilitate the verification.

    (4) The derived conditions are characterized by the form of algebraic inequalities, which are easy to put into effect in reality. The validity of theoretical results is confirmed via simulation results.

    Subsequently, the basic concepts, lemmas and the Caputo-type delayed ICGNNs model are described.

    Definition 1. [1] For a function u(), its fractional integral with order β is

    t0Iβtu(t)=1Γ(β)tt0(ts)β1u(s)ds,β>0.

    Definition 2. [39] For u()Cn([0,+],R), its Caputo derivative of the β-order is

    ct0Dβtu(t)=1Γ(nβ)tt0u(n)(s)(ts)βn+1ds,tt0.

    Lemma 1. [40] If ϕ(t) is differentiable and ϕ(t)C[t0,+), then

    ct0Dβt(ϕ2(t))2ϕ(t)ct0Dβtϕ(t),0<β<1.

    Lemma 2. [41] If ε>0, σ1,σ2R, then

    σ1σ2ε2σ21+12εσ22.

    Lemma 3. [42] For V(t)C[0,+), if satisfies c0DβtV(t)ΛV(t), then

    V(t)V(0)Eβ(Λtβ),

    where β(0,1) and ΛR.

    The Caputo-type delayed FOICGNNs is investigated as follows:

    c0D2βtrκ(t)=ˆδκc0Dβtrκ(t)ακ(rκ(t)){hκ(rκ(t))n=1dκψ(r(t))n=1bκψ(r(tτ1))n=1cκttτ2ψ(r(s))dsIκ(t)}, (2.1)

    where ακ(rκ(t)) is the amplification function; hκ(rκ(t)) means well behaved function; rκ(t) stands for the state of the k th neuron, ψ(r(t)) refers to the activation function, τ1 denotes the constant time delay, τ2 is the distributed delay; dκ,bκ and cκ are connection weights, Iκ(t) represents the external input and 0<|Iκ(t)|Iκ.

    Considering the following variable transform:

    wκ(t)=c0Dβtrκ(t)+πκrκ(t), (2.2)

    then, the system (2.1) is expressed as:

    c0Dβtrκ(t)=πκrκ(t)+wκ(t),c0Dβtwκ(t)=πκ(πκˆδκ)rκ(t)(ˆδκπκ)wκ(t)ακ(rκ(t))[hκ(rκ(t))n=1dκψ(r(t))n=1bκψ(r(tτ1))n=1cκttτ2ψ(r(s))dsIκ(t)]. (2.3)

    Analogously, the response system of the system (2.3) can be depicted as:

    c0Dβtyκ(t)=πκyκ(t)+vκ(t)+Pκ(t),c0Dβtvκ(t)=πκ(πκˆδκ)yκ(t)(ˆδκπκ)vκ(t)ακ(yκ(t))[hκ(yκ(t))n=1dκψ(y(t))n=1bκψ(y(tτ1))n=1cκttτ2ψ(y(s))dsIκ(t)]+Qκ(t). (2.4)

    We set the initial-value of systems (2.3) and (2.4) as:

    rκ(0)=rκ(0),wκ(t)=wκ(0);yκ(0)=yκ(0),vκ(0)=vκ(0).

    Let ˜eκ(t)=yκ(t)rκ(t), ˜zκ(t)=vκ(t)wκ(t), the above error systems are given by:

    c0Dβt˜eκ(t)=πκ˜eκ(t)+˜zκ(t)+Pκ(t),c0Dβt˜zκ(t)=Qκ(t)πκ(πκˆδκ)˜eκ(t)(ˆδκπκ)˜zκ(t)[ακ(yκ(t))hκ(yκ(t))ακ(rκ(t))hκ(rκ(t))]+ακ(yκ(t))[n=1dκ(ψ(y(t))ψ(r(t)))+n=1bκ(ψ(y(tτ1))ψ(r(tτ1)))]+ακ(rκ(t))n=1cκ[ttτ2ψ(y(s))dsttτ2ψ(r(s))ds]+[ακ(yκ(t))ακ(rκ(t))][n=1dκψ(r(t))+n=1bκψ(r(tτ1))+n=1cκttτ2ψ(y(s))ds+Iκ(t)]. (2.5)

    Assumption 1. The function of α() satisfies:

    0α_ακ(rκ(t))ˉα<,|ακ(y)ακ(r)|ακ|yr|,

    where y,rR and ακ>0.

    Assumption 2. If exists ϖκ>0, the functions α() and h() satisfy:

    ακ(r)hκ(r)ακ(y)hκ(y)ryϖκ,

    where ry,r,yR.

    Assumption 3. y,rR, if the activation function ψ() satisfies:

    |ψ(y)ψ(r)|Ψ|yr|,

    where Ψ>0.

    Assumption 4. For ψ(), the constant M>0 exist such that

    |ψ()|M.

    Definition 3. The systems (2.3) and (2.4) can reach GMLS under the controllers Pκ(t) and Qκ(t), if there exists one constant >0, such that

    ˜e(t)2+˜z(t)2{(˜e(0)22+˜z(0)22)Eβ(Λtβ)}12,0<β<1.

    The controllers of system (2.5) are designed:

    {Pκ(t)=˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)2˜zκ(t)n=1ˉα|cκ|Mτ2˜eκ(t),μκ=ˉα|bκ|Ψκϵ12;Qκ(t)=˘ϱκ˜zκ(t). (3.1)

    where ˜eκ(t)0.

    Theorem 1. The systems (2.3) and (2.4) can realize GMLS, when Λ>0,Λ=min{Λ1,Λ2}

    Λ1=πκ+˘ηκ[1+ϖκ+|π2κπκˆδκ|+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψκ)+ακIκ]ϵ12;Λ2=˘ϱκπκ+ˆδκ[1+|π2κπκˆδκ|+ϖκ+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψ+ˉα|bκ|Ψ)+ακIκ]12ϵ1.

    Proof. Constructing V() as:

    V(t)=12nκ=1[˜e2κ(t)+˜z2κ(t)]. (3.2)

    Based on Lemma 1, one has:

    c0DβtV(t)nκ=1˜eκ(t)c0Dβt˜eκ(t)+nκ=1˜zκ(t)c0Dβt˜zκ(t). (3.3)

    Substituting (2.5) into (3.3), we have

    c0DβtV(t)nκ=1˜eκ(t)[πκ˜eκ(t)+˜zκ(t)+Pκ(t)]+nκ=1˜zκ(t){πκ(πκˆδκ)˜eκ(t)(ˆδκπκ)˜zκ(t)[ακ(yκ(t))hκ(yκ(t))ακ(rκ(t))hκ(rκ(t))]+ακ(yκ(t))[n=1dκ(ψ(y(t))ψ(r(t)))+n=1bκ(ψ(y(tτ1))ψ(r(tτ1)))]+ακ(rκ(t))n=1cκttτ2ψ(y(s))dsακ(rκ(t))n=1cκttτ2ψ(r(s))ds+[ακ(yκ(t))ακ(rκ(t))][n=1dκψ(r(t))+n=1bκψ(r(tτ1))+n=1cκttτ2ψ(y(s))ds+Iκ(t)]+Qκ(t)}=nκ=1πκ˜e2κ(t)+nκ=1˜eκ(t)˜zκ(t)+nκ=1˜eκ(t)Pκ(t)+nκ=1(π2κ+πκˆδκ)˜eκ(t)˜zκ(t)+nκ=1(πκˆδκ)˜z2κ(t)nκ=1˜zκ(t)[ακ(yκ(t))hκ(yκ(t))ακ(rκ(t))hκ(rκ(t))]+nκ=1˜zκ(t)ακ(yκ(t))[n=1dκ(ψ(y(t))ψ(r(t)))]+nκ=1˜zκ(t)ακ(yκ(t))[n=1bκ(ψ(y(tτ1))ψ(r(tτ1)))]+nκ=1˜zκ(t)ακ(rκ(t))[n=1cκttτ2ψ(y(s))dsn=1cκttτ2ψ(r(s))ds]+nκ=1˜zκ(t){[ακ(yκ(t))ακ(rκ(t))][n=1dκψ(r(t))+n=1bκψ(r(tτ1))+n=1cκttτ2ψ(y(s))ds+Iκ(t)]+Qκ(t)}. (3.4)

    From Assumptions 1 and 2, it gives

    c0DβtV(t)nκ=1πκ˜e2κ(t)+nκ=1˜eκ(t)˜zκ(t)+ni=1˜eκ(t)Pκ(t)+ni=1(π2κ+πκˆδκ)˜eκ(t)˜zκ(t)+nκ=1(πκˆδκ)˜z2κ(t)+nκ=1˜zκ(t)ˉαn=1dκ[(ψ(y(t))ψ(r(t)))]+nκ=1˜zκ(t)ˉαn=1bκ[(ψ(y(tτ1))ψ(r(tτ1)))]nκ=1˜zκ(t)ϖκ˜eκ(t)+nκ=1˜zκ(t)n=1cκˉα[ttτ2ψ(y(s))dsttτ2ψ(r(s))ds]+nκ=1˜zκ(t){ακ˜eκ(t)[n=1dκψ(r(t))+n=1bκψ(r(tτ1))+n=1cκttτ2ψ(y(s))ds+Iκ(t)]}+nκ=1˜zκ(t)Qκ(t). (3.5)

    By Assumptions 3 and 4, one obtains

    c0DβtV(t)nκ=1πκ˜e2κ(t)+nκ=1˜eκ(t)˜zκ(t)+nκ=1[˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)2˜zκ(t)n=1ˉα|cκ|Mτ2]+nκ=1(π2κ+πκˆδκ)˜eκ(t)˜zκ(t)+nκ=1(πκˆδκ)˜z2κ(t)nκ=1ϖκ˜zκ(t)˜eκ(t)+nκ=1n=1ˉαdκΨ|e(t)|˜zκ(t)+nκ=1n=1ˉαbκΨ|e(tτ1)|˜zκ(t)+2nκ=1˜zκ(t)n=1ˉα|cκ|Mτ2+nκ=1˜zκ(t)˜eκ(t)[n=1ακ|dκ|M+nκ=1ακ|bκ|M+n=1ακ|cκ|Mτ2+ακIκ(t)]nκ=1˘ϱκ˜z2κ(t). (3.6)

    From Lemma 2, one has

    ˜eκ(t)˜zκ(t)ϵ12˜e2κ(t)+12ϵ1˜z2κ(t), (3.7)
    ˜eκ(t)˜eκ(tτ1)ϵ12˜e2κ(t)+12ϵ1˜e2κ(tτ1), (3.8)
    ˜e(t)˜zκ(t)ϵ12˜e2(t)+12ϵ1˜z2κ(t), (3.9)
    ˜e(tτ1)˜zκ(t)ϵ12˜e2(tτ1)+12ϵ1˜z2κ(t). (3.10)

    Substituting (3.7)–(3.10) into (3.6), then

    c0DβtV(t)nκ=1πκ˜e2κ(t)+nκ=1[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1[˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)]+nκ=1|π2κπκˆδκ|[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1(πκˆδκ)˜z2κ(t)+nκ=1ϖκ[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1n=1ˉα|dκ|Ψ[ϵ12˜e2(t)+12ϵ1˜z2κ(t)]+nκ=1n=1ˉα|bκ|Ψ[ϵ12˜e2(tτ1)+12ϵ1˜z2κ(t)]+nκ=1[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)][n=1ακ|dκ|M+n=1ακ|bκ|M+n=1ακ|cκ|Mτ2+ακIκ(t)]nκ=1˘ϱκ˜z2κ(t)nκ=1˜e2κ(t){πκ˘ηκ+[1+ϖκ+|π2κπκˆδκ|+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψκ)+ακIκ]ϵ12}+nκ=1˜z2κ(t){˘ϱκ+(πκˆδκ)+[1+|π2κπκˆδκ|+ϖκ+n=1(ακ|dκ|M+ακ|bκ|M+ακ|cκ|Mτ2+ˉα|dκ|Ψ+ˉα|bκ|Ψ)+ακIκ]12ϵ1}+nκ=1n=1˜e2κ(tτ1)(ˉα|bκ|Ψκϵ12μκ)=nκ=1˜e2κ(t){πκ+˘ηκ[1+ϖκ+|π2κπκˆδκ|+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψκ)+ακIκ]ϵ12}nκ=1˜z2κ(t){˘ϱκπκ+ˆδκ[1+|π2κπκˆδκ|+ϖκ+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψ+ˉα|bκ|Ψ)+ακIκ]12ϵ1}ΛV(t). (3.11)

    Because of the condition Λ=min{Λ1,Λ2}>0, where,

    Λ1=πκ+˘ηκ[1+ϖκ+|π2κπκˆδκ|+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψκ)+ακIκ]ϵ12;Λ2=˘ϱκπκ+ˆδκ[1+|π2κπκˆδκ|+ϖκ+n=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψ+ˉα|bκ|Ψ)+ακIκ]12ϵ1.

    From Lemma 3, one can get

    V(t)V(0)Eβ(Λtβ). (3.12)

    Therefore,

    ˜e(t)2+˜z(t)2[2nκ=1˜e2κ(t)+2nκ=1˜z2κ(t)]12=[4(12nκ=1˜e2κ(t)+12nκ=1˜z2κ(t))]12[4(V(0)Eβ(Λtβ)]12={2[˜e(0)22+˜z(0)22]Eβ(Λtβ)}12. (3.13)

    By Definition 3, the systems (2.3) and (2.4) can realize the GMLS.

    In order to realize the automatic adjustment of control gains, the controller of system (2.5) are designed as follows:

    {Pκ(t)=˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)2˜zκ(t)ˉαn=1|cκ|Mτ2˜eκ(t),μκ=ϵ12ˉα|bκ|Ψκ;Qκ(t)=ϑκ(t)˜zκ(t),Dβtϑκ(t)=ρκ˜z2κ(t), (3.14)

    where ˜eκ(t)0.

    Theorem 2. If ϑκ satisfies Λ>0, when Λ=min{Λ1,Λ2},

    Λ1=ϵ12[1+|πκ(πκˆδκ)|+ϖκ+ˉα|dκ|Ψκ+n=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]+πκ+˘ηκ,Λ2=12ϵ1[1+|πκ(πκˆδκ)|+ϖκ+ˉα(|dκ|+|bκ|)Ψ+n=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]πκ+ˆδκ+ϑκ,

    the systems (2.3) and (2.4) can realize GMLS.

    Proof. Designing V() is

    V(t)=12nκ=1˜e2κ(t)+12nκ=1˜z2κ(t)+nκ=112ρκ[ϑκ(t)ϑκ]2. (3.15)

    By utilizing Lemma 1, the Caputo type derivative of (3.13) can be obtained,

    c0DβtV(t)nκ=1˜eκ(t)c0Dβt˜eκ(t)+nκ=1˜zκ(t)c0Dβt˜zκ(t)+nκ=1[ϑκ(t)ϑκ]˜z2κ(t). (3.16)

    Substituting (2.5) into (3.16) yields

    c0DβtV(t)nκ=1˜eκ(t)[πκ˜eκ(t)+˜zκ(t)+Pκ(t)]+nκ=1˜zκ(t){πκ(πκˆδκ)˜eκ(t)(ˆδκπκ)˜zκ(t)[ακ(yκ(t))hκ(yκ(t))ακ(rκ(t))hκ(rκ(t))]+ακ(yκ(t))[n=1dκ(ψ(y(t))ψ(r(t)))+n=1bκ(ψ(y(tτ1))ψ(r(tτ1)))]+ακ(rκ(t))n=1cκttτ2ψ(y(s))dsακ(rκ(t))n=1cκttτ2ψ(r(s))ds+[ακ(yκ(t))ακ(rκ(t))][n=1dκψ(r(t))+n=1bκψ(r(tτ1))+n=1cκttτ2ψ(y(s))ds+Iκ(t)]+Qκ(t)}+nκ=1[ϑκ(t)ϑκ]˜z2κ(t). (3.17)

    From Assumptions 1 and 2, one has

    c0DβtV(t)nκ=1πκ˜e2κ(t)+nκ=1˜eκ(t)˜zκ(t)+nκ=1[˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)2˜zκ(t)ˉαn=1cκMτ2]+nκ=1[πκ(πκˆδκ)]˜zκ(t)˜eκ(t)nκ=1(ˆδκπκ)˜z2κ(t)nκ=1ϖκ˜eκ(t)˜zκ(t)+nκ=1ˉα˜zκ(t)[n=1dκ(ψ(y(t))ψ(r(t)))+n=1bκ(ψ(y(tτ1))ψ(r(tτ1)))]+nκ=1ˉαn=1cκ˜zκ(t)[ttτ2ψ(y(s))dsttτ2ψ(r(s))ds]+nκ=1ακ˜eκ(t)˜zκ(t)[n=1dκψ(r(t))+n=1bκψ(r(tτ1))+n=1cκttτ2ψ(y(s))ds+Iκ]nκ=1ϑκ(t)˜z2κ(t)+nκ=1ϑκ(t)˜z2κ(t)nκ=1ϑκ˜z2κ(t). (3.18)

    According to Assumptions 3 and 4, we get

    c0DβtV(t)nκ=1πκ˜e2κ(t)+nκ=1˜eκ(t)˜zκ(t)+nκ=1[˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)2nκ=1˜zκ(t)ˉαn=1|cκ|Mτ2]+nκ=1[πκ(πκˆδκ)]˜zκ(t)˜eκ(t)nκ=1(ˆδκπκ)˜z2κ(t)nκ=1ϖκ˜eκ(t)˜zκ(t)+nκ=1ˉα˜zκ(t)[n=1dκΨ|˜e(t)|+n=1bκΨ|˜e(tτ1)|]+2nκ=1ˉα˜zκ(t)n=1|cκ|Mτ2+nκ=1ακ˜eκ(t)˜zκ(t)[n=1|dκ|M+n=1|bκ|M+n=1|cκ|Mτ2+Iκ]nκ=1ϑκ˜z2κ(t). (3.19)

    From (3.7)–(3.10), then we get

    c0DβtV(t)nκ=1πκ˜e2κ(t)+nκ=1[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1[˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)]+nκ=1|πκ(πκˆδκ)|[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]nκ=1(ˆδκπκ)˜z2κ(t)+nκ=1ϖκ[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1n=1ˉα|dκ|Ψ[ϵ12˜e2(t)+12ϵ1˜z2κ(t)]+nκ=1n=1ˉα|bκ|Ψ[ϵ12˜e2(tτ1)+12ϵ1˜z2κ(t)]+nκ=1ακ[n=1|dκ|M+n=1|bκ|M+n=1|cκ|Mτ2+Iκ][ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]nκ=1ϑκ˜z2κ(t)nκ=1πκ˜e2κ(t)+nκ=1[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1[˘ηκ˜e2κ(t)μκ˜e2κ(tτ1)]+nκ=1|πκ(πκˆδκ)|[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]nκ=1(ˆδκπκ)˜z2κ(t)+nκ=1ϖκ[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)]+nκ=1n=1ˉα|dκ|Ψκϵ12˜e2κ(t)+nκ=1n=1ˉα|dκ|Ψ12ϵ1˜z2κ(t)+nκ=1n=1ˉα|bκ|Ψκϵ12˜e2κ(tτ1)+nκ=1ακ[ϵ12˜e2κ(t)+12ϵ1˜z2κ(t)][n=1|dκ|M+n=1|bκ|M+n=1|cκ|Mτ2+Iκ]+nκ=1n=1ˉα|bκ|Ψ12ϵ1˜z2κ(t)nκ=1ϑκ˜z2κ(t)nκ=1{ϵ12[1+|πκ(πκˆδκ)|+ϖκ+ˉα|dκ|Ψκ+ακ(n=1|dκ|M+n=1|bκ|M+n=1|cκ|Mτ2+Iκ)]πκ˘ηκ}˜e2κ(t)+nκ=1{12ϵ1[1+|πκ(πκˆδκ)|+ϖκ+ˉα(|dκ|+|bκ|)Ψ+ακ(n=1|dκ|M+n=1|bκ|M+n=1|cκ|Mτ2+Iκ)]+πκˆδκϑκ}˜z2κ(t)+nκ=1(ϵ12ˉα|bκ|Ψκμκ)˜e2κ(tτ1)nκ=1˜e2κ(t){ϵ12[1+|πκ(πκˆδκ)|+ϖκ+ˉα|dκ|Ψκ+n=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]+πκ+˘ηκ}nκ=1˜z2κ(t){12ϵ1[1+|πκ(πκˆδκ)|+ϖκ+ˉα(|dκ|+|bκ|)Ψ+n=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]πκ+ˆδκ+ϑκ}ΛV(t). (3.20)

    Based on the condition Λ=min{Λ1,Λ2}>0, there,

    Λ1=ϵ12[1+|πκ(πκˆδκ)|+ϖκ+ˉα|dκ|Ψκ+n=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]+πκ+˘ηκ;Λ2=12ϵ1[1+|πκ(πκˆδκ)|+ϖκ+ˉα(|dκ|+|bκ|)Ψ+n=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]πκ+ˆδκ+ϑκ.

    From Lemma 3, then

    V(t)V(0)Eβ(Λtβ). (3.21)

    Therefore,

    ˜e(t)2+˜z(t)2[2nκ=1˜e2κ(t)+2nκ=1˜z2κ(t)]12=[4(12nκ=1˜e2κ(t)+12nκ=1˜z2κ(t))]12[4(V(0)Eβ(Λtβ)]12={2[e(0)22+z(0)22+nκ=11ρκ[ϑκ(0)ϑκ]2]Eβ(Λtβ)}12. (3.22)

    Note that,

    ˜e(0)22+˜z(0)22+nκ=11ρκ[ϑκ(0)ϑκ]2H(˜e(0)22+˜z(0)22), (3.23)

    where H is a positive constant.

    Thus,

    ˜e(t)2+˜z(t)2{2H(˜e(0)22+˜z(0)22)Eβ(Λtβ)}12. (3.24)

    From Definition 3, the systems (2.3) and (2.4) can realize GMLS.

    Remark 1. Different with the GAS [37], quasi projective synchronization [25], finite-time synchronization [26,35,36], Mean-square synchronization [38], fixed-time synchronization [7,26,28], exponential synchronization [29], pinning synchronization [43] and adaptive quantitative exponential synchronization [44]. The GMLS of ICGNNs is discussed in the article.

    Remark 2. Note that the attack resistant control [44], sliding mode control [14,45], pinning control [13,38] and intermittent control [46] are inconvenient and undesirable in practical applications. Two novel feedback controllers are designed to achieve GMLS in this article, which is more convenient. Different with the linear control and adaptive control, in this paper, two control inputs Pκ(t) and Qκ(t) are incorporated into the the model of response systems. It can be seen that the selection of the novel feedback controller with time delays is more efficient and extensive.

    Remark 3. Compared with the models of integer order ICGNNs considered in some literature [8,24,27,28,29], we study the FOICGNNs. Different with the dynamic properties exhibited by integer-order ICGNNs, FOICGNNs possess infinite memory characteristics and offer a more accurate depiction of neuronal dynamics. Consequently, they are particularly well-suited for application in realistic domains. Different from the model [30], the model of this article is based on the derivative of Caputo' sense. In other researches, various types of time delays have been considered such as proportional delay [27,31], discrete delay [30,47], time-varying delay [13,33] and leakage delay [47]. However, the system of this article are considered both discrete and distributed delays. For these reasons, the proposed model of the nervous system is less conservative.

    Remark 4. Different from some approaches such as event-triggered approach [28], matrix inequalities approach [33] and contraction mapping principle [32], several efficient guidelines are derived to implement the GMLS of Caputo FOICGNNs through the method of variable substitution and inequality technique. The resulting criterion is formulated in the form of an algebraic inequality, which greatly reduces the computational complexity and is easier to prove and implement in practical applications.

    Remark 5. In [37], the GAS of fractional-order dynamic systems was realized within the framework of event-based delayed impulsive control. However, in this paper, the GMLS of Caputo-type delayed ICGNNs is considered. As a matter of fact, the GMLS is simplified to the GAS. Thus, the GMLS of this paper can apply to the GAS. In [43], pinning synchronization of integer-order NNs was achieved. Instead, the model of fractional-order is discussed in this paper, which can overcome the disadvantage that the integer order model does not agree well with the experimental results. In addition, the fractional-order model is more clear and concise in describing complex physical and mechanical problems.

    Some simulations are provided to check the availability of the results in this part.

    Example 1. Considering two-dimensional FOICGNNs with discrete and distributed delays:

    c0D2βtrκ(t)=ˆδκc0Dβtrκ(t)ακ(rκ(t)){hκ(rκ(t))2=1dκψ(r(t))2=1bκψ(r(tτ1))2=1cκttτ2ψ(r(s))dsIκ(t)}, (4.1)

    where κ=1,2, and 0<β<1, let: β=0.95, α1(r1)=0.9+0.21+r12, α2(r2)=0.9+0.21+r22, Iκ(t)=sint, ψ()=cos(), h1(r1)=0.2r10.21+r12, h2(r2)=0.2r20.21+r22. By calculating α_=0.9, ¯α=1.1, α1=α2=0.2, ϖ1=ϖ2=0.18.

    Then, choose π1=2.9, π2=2.9, ˆδ1=4.6, ˆδ2=4.6  d11=0.23,  d12=1.5,  d21=0.64,  d22=0.3,  b11=0.12,  b12=0.1,  b21=0.2,  b22=0.16,  c11=0.15,  c12=0.3,  c21=0.2,  c22=0.15, τ1=τ2=1. ϵ1=1. The Lipchitz constants Ψ1=Ψ2=0.5,  ψ()=cos().  I1=I2=1.

    The parameters of the controllers, we choose:  ˘η1=8.6,  ˘η2=9.1,  ˘ϱ1=9.0,  ˘ϱ2=8.7,  μ1=0.088,  μ2=0.0715,  2ˉα(c11+c12)=0.99, 2ˉα(c21+c22)=0.77. The initial values are given as r1(0)=2.9, r2(0)=1.5, w1(0)=5.7, w2(0)=1.5, y1(0)=0.4, y2(0)=2.3, v1(0)=1.9, v2(0)=0.1. Therefore,

    Λ1=πκ+˘ηκ[1+ϖκ+|π2κπκˆδκ|+2=1(ακM(|dκ|+|bκ|+|cκ|τ2)+ˉα|dκ|Ψκ)+ακIκ]ϵ12;Λ2=˘ϱκπκ+ˆδκ[1+|π2κπκˆδκ|+ϖκ+2=1(ακM(|dκ|+|bκ|+cκ|τ2)+ˉα|dκ|Ψ+ˉα|bκ|Ψ)+ακIκ]12ϵ1,

    when κ=1, Λ1=7.86575, Λ2=6.76875; κ=2, Λ1=8.185, Λ2=6.7225; then Λ=min{Λ1,Λ2}=6.7225>0. Thus, the systems (2.3) and (2.4) can realize the GMLS. The figures of each state variable of the systems (2.3) and (2.4) are given in Figure 1. As is shown in the picture, the curves for the state of the system are eventually coeval. The error curves of systems (2.3) and (2.4) are portrayed in Figure 2.

    Figure 1.  The state curves of systems (2.3) and (2.4).
    Figure 2.  Evolution of synchronization error of systems (2.3) and (2.4).

    Example 2. Considering two-dimensional models (2.3) and (2.4): Where κ=1,2, and 0<β<1, let: β=0.9, α1(r1)=0.3+0.61+r12, α2(r2)=0.3+0.61+r22, Iκ(t)=sint, ψ()=cos(), I1=I2=1, h1(r1)=cos(r1)0.41+r12, h2(r2)=cos(r2)0.41+r22. By calculating α_=0.3, ¯α=0.9, α1=α2=0.6, ϖ1=ϖ2=0.3.

    The relevant parameters are selected as π1=3.4, π2=3.4, ˆδ1=3.9, ˆδ2=3.9  d11=0.3,  d12=1.4,  d21=0.03,  d22=1.0,  b11=0.12,  b12=0.1,  b21=0.18,  b22=0.13,  c11=0.45,  c12=0.14,  c21=0.2,  c22=0.4, τ1=1, τ2=1. The Lipchitz constants Ψ1=Ψ2=0.5. Taking the initial values as: r1(0)=2.9, r2(0)=3.5, w1(0)=2.5, w2(0)=1.9, y1(0)=0.8, y2(0)=2.4, v1(0)=1.9, v2(0)=1.5.

    Under the controller (3.14), select ρ1=1, ρ2=1.5, ϑ1(0)=1.2, ϑ2(0)=1.8, ϑ1=7, ϑ2=5,  ˘η1=7.0,  ˘η2=7.4,  μ1=0.0675,  μ2=0.05175, 2ˉα(c11+c12)=1.062, 2ˉα(c21+c22)=1.08, Therefore,

    Λ1=ϵ12[1+|πκ(πκˆδκ)|+ϖκ+ˉα|dκ|Ψκ+2=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]+πκ+˘ηκ;Λ2=12ϵ1[1+|πκ(πκˆδκ)|+ϖκ+ˉα(|dκ|+|bκ|)Ψ+2=1ακM(|dκ|+|bκ|+|cκ|τ2)+ακIκ]πκ+ˆδκ+ϑκ,

    when κ=1, Λ1=7.76375, Λ2=4.515; κ=2, Λ1=7.878, Λ2=2.8165; then Λ=min{Λ1,Λ2}=6.7225>0.

    Therefore, the simulations confirm the validity of proposed Theorem 2, the systems (2.3) and (2.4) can reach GMLS. The curves of control gains are provided in Figure 5. In Figure 3 represents the state trajectories of the systems (2.3) and (2.4). The synchronization error of systems (2.3) and (2.4) are exhibited in Figure 4.

    Figure 3.  Trajectories state of systems (2.3) and (2.4).
    Figure 4.  Synchronization error of systems (2.3) and (2.4).
    Figure 5.  Evolution of adaptive gains ϑ1(t) and ϑ2(t).

    The GMLS analysis of delayed FOICGNNs is presented in this article. By means of variable substitution and Caputo-type derivative property, the 2β order inertial system is converted into two general systems with β order. Some criteria are formulated by the use of the differential mean-value theorem and fractional differential inequality techniques, the GMLS of FOICGNNs is implemented by designing the novel controllers. Ultimately, two numerical simulations are furnished to demonstrate the feasibility of the results. In our future works, we will further investigate event-triggered control on finite-time Mittag-Leffler synchronization of FOICGNNs with time-varying delays.

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

    This work was supported by the Natural Science Foundation of Anhui Province of China (No.1908085MA01), the Top Young Talents Program of Higher Learning Institutions of Anhui Province of China (No.gxyq2019048) and the Natural Science Foundation of the Higher Education Institutions of Anhui Province of China (No.2023AH050478).

    The authors declare that they have no competing interests.



    [1] I. Podlubny, Fractional differential equations, New York: Academic Press, 1999.
    [2] R. Koeller, Applications of fractional calculus to the theory of viscoelasticity, J. Appl. Mech., 51 (1984), 229–307. https://doi.org/10.1115/1.3167616 doi: 10.1115/1.3167616
    [3] D. Baleanu, J. Machado, A. Luo, Fractional dynamics and control, Berlin: Springer, 2012.
    [4] X. Yang, C. Li, T. Huang, Q. Song, Mittag-Leffler stability analysis of nonlinear fractional-order systems with impulses, Appl. Math. Comput., 293 (2017), 416–422. https://doi.org/10.1016/j.amc.2016.08.039 doi: 10.1016/j.amc.2016.08.039
    [5] F. Hoppensteadt, E. Izhikevich, Pattern recognition via synchronization in phase-locked loop neural networks, IEEE T. Neural Networ., 11 (2000), 734–738. https://doi.org/10.1109/72.846744 doi: 10.1109/72.846744
    [6] F. Du, J. Lu, New criteria on finite-time stability of fractional-order hopfield neural networks with time delays, IEEE T. Neur. Net. Lear., 32 (2021), 3858–3866. https://doi.org/10.1109/TNNLS.2020.3016038 doi: 10.1109/TNNLS.2020.3016038
    [7] L. Zhou, H. Lin, F. Tan, Fixed/predefined-time synchronization of coupled memristor-based neural networks with stochastic disturbance, Chaos Soliton. Fract., 173 (2023), 113643. https://doi.org/10.1016/j.chaos.2023.113643 doi: 10.1016/j.chaos.2023.113643
    [8] S. Han, C. Hu, J. Yu, H. Jiang, S. Wen, Stabilization of inertial Cohen-Grossberg neural networks with generalized delays: A direct analysis approach, Chaos Soliton. Fract., 142 (2021), 110432. https://doi.org/10.1016/j.chaos.2020.110432 doi: 10.1016/j.chaos.2020.110432
    [9] C. Long, G. Zhang, Z. Zeng, J. Hu, Finite-time stabilization of complex-valued neural networks with proportional delays and inertial terms: A non-separation approach, Neural Networks, 148 (2022), 86–95. https://doi.org/10.1016/j.neunet.2022.01.005 doi: 10.1016/j.neunet.2022.01.005
    [10] N. Cui, H. Jiang, C. Hu, A. Abdurahman, Global asymptotic and robust stability of inertial neural networks with proportional delays, Neurocomputing, 272 (2018), 326–333. https://doi.org/10.1016/j.neucom.2017.07.001 doi: 10.1016/j.neucom.2017.07.001
    [11] J. Liu, L. Shu, Q. Chen, S. Zhong, Fixed-time synchronization criteria of fuzzy inertial neural networks via Lyapunov functions with indefinite derivatives and its application to image encryption, Fuzzy Set. Syst., 459 (2023), 22–42. https://doi.org/10.1016/j.fss.2022.08.002 doi: 10.1016/j.fss.2022.08.002
    [12] J. Han, G. Chen, J. Hu, New results on anti-synchronization in predefined-time for a class of fuzzy inertial neural networks with mixed time delays, Neurocomputing, 495 (2022), 26–36. https://doi.org/10.1016/j.neucom.2022.04.120 doi: 10.1016/j.neucom.2022.04.120
    [13] Y. Yu, Z. Zhang, M. Zhong, Z. Wang, Pinning synchronization and adaptive synchronization of complex-valued inertial neural networks with time-varying delays in fixed-time interval, J. Franklin I., 359 (2022), 1434–1456. https://doi.org/10.1016/j.jfranklin.2021.11.036 doi: 10.1016/j.jfranklin.2021.11.036
    [14] R. Guo, S. Xu, Observer-based sliding mode synchronization control of complex-valued neural networks with inertial term and mixed time-varying delays, Appl. Math. Comput., 442 (2023), 127761. https://doi.org/10.1016/j.amc.2022.127761 doi: 10.1016/j.amc.2022.127761
    [15] L. Ke, Mittag-Leffler stability and asymptotic ω-periodicity of fractional-order inertial neural networks with time-delays, Neurocomputing, 465 (2021), 53–62. https://doi.org/10.1016/j.neucom.2021.08.121 doi: 10.1016/j.neucom.2021.08.121
    [16] Y. Gu, H. Wang, Y. Yu, Stability and synchronization for Riemann-Liouville fractional-order time-delayed inertial neural networks, Neurocomputing, 340 (2019), 270–280. https://doi.org/10.1016/j.neucom.2019.03.005 doi: 10.1016/j.neucom.2019.03.005
    [17] Y. Cheng, H. Zhang, W. Zhang, H. Zhang, Novel algebraic criteria on global Mittag-Leffler synchronization for FOINNs with the Caputo derivative and delay, J. Appl. Math. Comput., 68 (2022), 3527–3544. https://doi.org/10.1007/s12190-021-01672-0 doi: 10.1007/s12190-021-01672-0
    [18] Q. Peng, J. Jian, Synchronization analysis of fractional-order inertial-type neural networks with time delays, Math. Comput. Simul., 205 (2022), 62–77. https://doi.org/10.1016/j.matcom.2022.09.023 doi: 10.1016/j.matcom.2022.09.023
    [19] A. Alimi, C. Aouiti, E. Assali, Finite-time and fixed-time synchronization of a class of inertial neural networks with multi-proportional delays and its application to secure communication, Neurocomputing, 332 (2019), 29–43. https://doi.org/10.1016/j.neucom.2018.11.020 doi: 10.1016/j.neucom.2018.11.020
    [20] M. Prakash, P. Balasubramaniam, S. Lakshmanan, Synchronization of Markovian jumping inertial neural networks and its applications in image encryption, Neural Networks, 83 (2016), 86–93. https://doi.org/10.1016/j.neunet.2016.07.001 doi: 10.1016/j.neunet.2016.07.001
    [21] J. Feng, J. Ma, S. Qin, Exponential stability of periodic solution for impulsive memristor-based cohen-grossberg neural networks with mixed delays, Int. J. Pattern Recogn., 31 (2017), 1750022. https://doi.org/10.1142/S0218001417500227 doi: 10.1142/S0218001417500227
    [22] C. Zhou, H. Zhang, H. Zhang, C. Dang, Global exponential stability of impulsive fuzzy Cohen-Grossberg neural networks with mixed delays and reaction Cdiffusion terms, Neurocomputing, 91 (2012), 67–76. https://doi.org/10.1016/j.neucom.2012.02.012 doi: 10.1016/j.neucom.2012.02.012
    [23] Z. Wang, H. Zhang, W. Yu, Robust stability of Cohen-Grossberg neural networks via state transmission matrix, IEEE T. Neural Networ., 20 (2009), 169–174. https://doi.org/10.1109/tnn.2008.2009119 doi: 10.1109/tnn.2008.2009119
    [24] Y. Li, J. Xiang, Existence and global exponential stability of anti-periodic solution for clifford-valued inertial Cohen-Grossberg neural networks with delays, Neurocomputing, 332 (2019), 259–269. https://doi.org/10.1016/j.neucom.2018.12.064 doi: 10.1016/j.neucom.2018.12.064
    [25] S. Baluni, V. Yadav, S. Das, Quasi projective synchronization of time varying delayed complex valued Cohen-Grossberg neural networks, Inform. Sciences, 612 (2022), 231–240. https://doi.org/10.1016/j.ins.2022.08.106 doi: 10.1016/j.ins.2022.08.106
    [26] F. Kong, Q. Zhu, R. Sakthivel, Finite-time and fixed-time synchronization analysis of fuzzy Cohen-Grossberg neural networks with discontinuous activations and parameter uncertainties, Eur. J. Control, 56 (2022), 179–190. https://doi.org/10.1016/j.ejcon.2020.03.003 doi: 10.1016/j.ejcon.2020.03.003
    [27] Q. Li, L. Zhou, Global asymptotic synchronization of inertial memristive Cohen-Grossberg neural networks with proportional delays, Commun. Nonlinear Sci., 123 (2023), 107295. https://doi.org/10.1016/j.cnsns.2023.107295 doi: 10.1016/j.cnsns.2023.107295
    [28] H. Jia, D. Luo, J. Wang, H. Shen, Fixed-time synchronization for inertial Cohen-Grossberg delayed neural networks: An event-triggered approach, Knowl.-Based Syst., 250 (2022), 109104. https://doi.org/10.1016/j.knosys.2022.109104 doi: 10.1016/j.knosys.2022.109104
    [29] L. Ke, W. Li, Exponential synchronization in inertial Cohen-Grossberg neural networks with time delays, J. Franklin I., 356 (2019), 11285–11304. https://doi.org/10.1016/j.jfranklin.2019.07.027 doi: 10.1016/j.jfranklin.2019.07.027
    [30] Z. Li, Y. Zhang, The boundedness and the global Mittag-Leffler synchronization of fractional-order inertial Cohen-Grossberg neural networks with time delays, Neural Process. Lett., 54 (2022), 597–611. https://doi.org/10.1007/s11063-021-10648-x doi: 10.1007/s11063-021-10648-x
    [31] H. Zhang, C. Wang, R. Ye, I. Stamova, J. Cao, Novel order-dependent passivity conditions of fractional generalized Cohen-Grossberg neural networks with proportional delays, Commun. Nonlinear Sci., 120 (2023), 107155. https://doi.org/10.1016/j.cnsns.2023.107155 doi: 10.1016/j.cnsns.2023.107155
    [32] L. Wan, Z. Liu, Multiple O(tq) stability and instability of time-varying delayed fractional-order Cohen-Grossberg neural networks with Gaussian activation functions, Neurocomputing, 454 (2021), 212–227. https://doi.org/10.1016/j.neucom.2021.05.018 doi: 10.1016/j.neucom.2021.05.018
    [33] R. Aravind, P. Balasubramaniam, Stability criteria for memristor-based delayed fractional-order Cohen-Grossberg neural networks with uncertainties, J. Comput. Appl. Math., 420 (2023), 114764. https://doi.org/10.1016/j.cam.2022.114764 doi: 10.1016/j.cam.2022.114764
    [34] H. Zhang, X. Chen, R. Ye, I. Stamova, J. Cao, Adaptive quasi-synchronization analysis for Caputo delayed Cohen-Grossberg neural networks, Math. Comput. Simulat., 212 (2023), 49–65. https://doi.org/10.1016/j.matcom.2023.04.025 doi: 10.1016/j.matcom.2023.04.025
    [35] J. Xiao, L. Wu, A. Wu, Z. Zeng, Z. Zhang, Novel controller design for finite-time synchronization of fractional-order memristive neural networks, Neurocomputing, 512 (2022), 494–502. https://doi.org/10.1016/j.neucom.2022.09.118 doi: 10.1016/j.neucom.2022.09.118
    [36] Z. Ruan, Y. Li, J. Hu, J. Mei, D. Xia, Finite-time synchronization of the drive-response networks by event-triggered aperiodic intermittent control, Neurocomputing, 485 (2022), 89–102. https://doi.org/10.1016/j.neucom.2022.02.037 doi: 10.1016/j.neucom.2022.02.037
    [37] B. Zheng, Z. Wang, Event-based delayed impulsive control for fractional-order dynamic systems with application to synchronization of fractional-order neural networks, Neural. Comput. Appl., 35 (2023), 20241–20251. https://doi.org/10.1007/s00521-023-08738-z doi: 10.1007/s00521-023-08738-z
    [38] X. Yuan, G. Ren, H. Wang, Y. Yu, Mean-square synchronization of fractional-order stochastic complex network via pinning control, Neurocomputing, 513 (2022), 153–164. https://doi.org/10.1016/j.neucom.2022.09.128 doi: 10.1016/j.neucom.2022.09.128
    [39] A. Kilbas, H. Srivastava, J. Trujillo, Theory and applications of fractional differential equations, Amsterdam: Elsevier, 2006.
    [40] M. D. Mermoud, N. A. Camacho, J. Gallegos, R. C. Linares, Using general quadratic Lyapunov functions to prove Lyapunov uniform stability for fractional order systems, Commun. Nonlinear Sci., 22 (2015), 650–659. https://doi.org/10.1016/j.cnsns.2014.10.008 doi: 10.1016/j.cnsns.2014.10.008
    [41] M. Tan, Asymptotic stability of nonlinear systems with unbounded delays, J. Math. Anal. Appl., 337 (2008), 1010–1021. https://doi.org/10.1016/j.jmaa.2007.04.019 doi: 10.1016/j.jmaa.2007.04.019
    [42] J. Chen, Z. Zeng, P. Jiang, Global Mittag-Leffler stability and synchronization of memristor-based fractional-order neural networks, Neural Networks, 51 (2014), 1–8. https://doi.org/10.1016/j.neunet.2013.11.016 doi: 10.1016/j.neunet.2013.11.016
    [43] L. Feng, C. Hu, J. Yu, H. Jiang, Pinning synchronization of directed networks with disconnected switching topology via averaging method, Nonlinear Anal.-Hybri., 49 (2023), 101369. https://doi.org/10.1016/j.nahs.2023.101369 doi: 10.1016/j.nahs.2023.101369
    [44] F. Tan, L. Zhou, J. Xia, Adaptive quantitative exponential synchronization in multiplex Cohen-Grossberg neural networks under deception attacks, J. Franklin I., 2022. https://doi.org/10.1016/j.jfranklin.2022.09.020. doi: 10.1016/j.jfranklin.2022.09.020
    [45] J. Xiao, Z. Zeng, A. Wu, S. Wen, Fixed-time synchronization of delayed Cohen-Grossberg neural networks based on a novel sliding mode, Neural Networks, 128 (2020), 1–12. https://doi.org/10.1016/j.neunet.2020.04.020 doi: 10.1016/j.neunet.2020.04.020
    [46] M. Hui, C. Wei, J. Zhang, H. Iu, R. Yao, L. Bai, Finite-time synchronization of fractional-order memristive neural networks via feedback and periodically intermittent control, Commun. Nonlinear Sci., 116 (2022), 160822. https://doi.org/10.1016/j.cnsns.2022.106822 doi: 10.1016/j.cnsns.2022.106822
    [47] H. Zhang, J. Cheng, H. Zhang, W. Zhang, J. Cao, Quasi-uniform synchronization of Caputo type fractional neural networks with leakage and discrete delays, Chaos Soliton. Fract., 152 (2021), 111432. https://doi.org/10.1016/j.chaos.2021.111432 doi: 10.1016/j.chaos.2021.111432
  • 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(1321) PDF downloads(76) Cited by(0)

Figures and Tables

Figures(5)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog