http://orcid.org/0000-0001-9949-8731
Figures
Abstract
Finite time synchronization, which means synchronization can be achieved in a settling time, is desirable in some practical applications. However, most of the published results on finite time synchronization don’t include delays or only include discrete delays. In view of the fact that distributed delays inevitably exist in neural networks, this paper aims to investigate the finite time synchronization of memristor-based Cohen-Grossberg neural networks (MCGNNs) with both discrete delay and distributed delay (mixed delays). By means of a simple feedback controller and novel finite time synchronization analysis methods, several new criteria are derived to ensure the finite time synchronization of MCGNNs with mixed delays. The obtained criteria are very concise and easy to verify. Numerical simulations are presented to demonstrate the effectiveness of our theoretical results.
Citation: Chen C, Li L, Peng H, Yang Y (2017) Finite time synchronization of memristor-based Cohen-Grossberg neural networks with mixed delays. PLoS ONE 12(9): e0185007. https://doi.org/10.1371/journal.pone.0185007
Editor: Jun Ma, Lanzhou University of Technology, CHINA
Received: June 2, 2017; Accepted: September 4, 2017; Published: September 20, 2017
Copyright: © 2017 Chen et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper.
Funding: The work is supported by the National Key Research and Development Program (Grant nos. 2016YFB0800602 and 2016YFB0800604), and the National Natural Science Foundation of China (Grant Nos. 61573067, 61771071 and 61472045).
Competing interests: The authors have declared that no competing interests exist.
Introduction
Memristor, which was first proposed by Chua in 1971 [1], is deemed as the fourth fundamental circuit element besides inductor, capacitor and resistor. In 2008, the prototype of memristor was first realized by the scientists of Hewlett-Packard (HP) [2]. Memristor, the contraction of memory resistor, reflects the nonlinear relationship between charge and flux (see Fig 1). It has been proved that memristor has variable resistance and the function of memory. In the artificial neural network, the synapses are usually modeled by resistors [3]. Since memristors own memory and perform more like real biological synapses, now memristors have been utilized to replace the resistors in artificial neural network to build memristor-based neural network (MNN), which is the appropriate candidate for simulating the human brain [4].
On the other hand, synchronization of complex networks [5–7] has received much attention due to its great application prospect in many different fields such as image encryption [8], secure communications [9] and associative memory [10]. By utilizing a memristor to replace the diode in Chuas circuit, Chua obtained several oscillators in [11]. Since then, various memristive chaotic systems have been proposed by using the similar methods. As we know, chaos [12, 13] presents complex nonlinear behaviours, but it can appears in a simple memristor-based Chuas circuit! So it is important to study the synchronization control of MNNs [14–17]. In [14], the exponential synchronization of MNNs with mixed delays was investigated via adaptive control. By means of intermittent control, the authors of [15] studied the stability and synchronization of memristor-based coupling neural networks with time-varying delays. Finite-time Mittag-Leffler synchronization of fractional-order memristive BAM neural networks with time delays was studied in [17]. Moreover, in view of the characteristics that signals transmit in real neural network, we should study the neural networks with time delays, including discrete delays [18, 19] and distributed delays [20, 21].
It should be pointed out that controller plays an important part in realizing synchronization. But how to design the optimal controller? It seems this problem has not been addressed. So far, many effective control methods have been proposed, such as the activation control [22], pinning control [23, 24], the linear separation method [25], the linear coupling method [26], impulsive control [27], adaptive control [14, 28], intermittent control [15], the sliding mode control [29], etc. However, in this paper, the controller that we design is a discontinuous feedback controller. Compared with the above-mentioned control methods, feedback control has the simplest form, and is very easy to be manipulated in practical applications.
In 1983, Cohen and Grossberg [30] proposed the Cohen-Grossberg neural network model, which is very general and important in all kinds of neural network models. Some important neural networks, such as cellular neural network and Hopfield neural network, can be deemed as the special cases of Cohen-Grossberg neural network. Although there have been many results about delayed Cohen-Grossberg neural networks [31–40], few of them are related to delayed MCGNNs. Recently, the exponential synchronization of MCGNNs with mixed delays was discussed in [41], the function projective synchronization of MCGNNs with time-varying discrete delays was studied in [42]. Obviously, most published results about synchronization control of MCGNNs only consider asymptotical synchronization and exponential synchronization, which mean that the synchronization time is infinite, but in application fields, it is more meaningful that the synchronization can be achieved in finite time. However, up to now, only Ref. [43] was concerning the finite time synchronization of MCGNNs with discrete delays. It should be pointed out that the distributed delays were not considered in [43], and the controller used in [43] was very complicated. As far as we know, there has been no published result on finite time synchronization of MCGNNs with mixed delays until now.
Inspired by the above analysis, this paper is devoted to studying the finite time synchronization problem of MCGNNs with mixed delays. The main contributions and originality of our paper are listed below: (i) This is the first attempt to investigate the finite time synchronization problem of MCGNNs with mixed delays, including time-varying discrete delays and distributed delays. Compared with the results in [43], the results in this paper are more general. (ii) The finite time synchronization analysis method used in this paper is a novel finite time synchronization analysis method, which has only been used in our another paper [44]. By adopting this novel analysis method, we derive some sufficient conditions that can ensure the finite time synchronization of the studied MCGNNs. Furthermore, the analysis method used in this paper can also be applied to analyze the finite time synchronization of other MNNs. (iii) In many literatures on the finite time synchronization of delayed systems, the controllers are very complicated. Although the neural network model considered in this paper is MCGNN with mixed delays, only simple feedback controllers are enough to derive the finite time synchronization of the studied MCGNNs. In some papers, the designed controllers were also similar to the controllers in this paper, however, only the asymptotical synchronization or the exponential synchronization of the studied systems can be obtained.
The rest of this paper is organized as follows. Some essential preliminaries are introduced in Section 2. In Section 3, our main results are derived. In Section 4, numerical simulations are presented to verify the theoretical results. Conclusions are drawn in Section 5.
Preliminaries
Referring to some existing MCGNN models [41–43], in this paper, we consider the following MCGNN with mixed delays:
(1)
where ξi(t) represents the state of the ith neuron; wi(⋅) is the amplification function; ai(⋅) denotes the appropriately behaved function; τij(t) is the discrete delay; Kij: [0, +∞) → [0, +∞) stands for the delay kernel of the unbounded distributed delay; Ii is the external input; the initial value of MCGNN (1) is φ(s) = (φ1(s), φ2(s), …, φn(s))T, s ≤ 0; bij(ξi(t)), cij(ξi(t)) and dij(ξi(t)) are memristive connection weights, which are given by
(2)
for i, j = 1, 2, …, n, where
are known constants [45, 46].
To derive the theoretical results, some assumptions will be needed:
(A1) 0 ≤ τij(t) ≤ τij, , where τij > 0 and σij > 0 are constants, i, j = 1, 2, …, n.
(A2) wi(⋅) is continuous and , where
and
are constants, i = 1, 2, …, n.
(A3) , where ai > 0 are constants, i = 1, 2, …, n.
(A4) For ∀x, y ∈ R, there exist constants li > 0 such that
(A5) There exist constants Mi > 0 such that |fi(⋅)| ≤ Mi, i = 1, 2, …, n.
(A6) There exist constants Kij > 0 such that
Choose a transformation function Φi(⋅), which satisfies
(3)
In view of
, we know Φi(⋅) is strictly monotone increasing, then
exists. Let xi(t) = Φi(ξi(t)), we have
. On the other hand, by the derivative theorem of inverse function,
. Then it follows that
(4)
Throughout this paper, we set ,
,
, for i, j = 1, 2, …, n.
stands for the closure of the convex hull generated by set E.
Based on the relevant theories of differential inclusions and set-valued maps [47, 48], we can derive that:
(5)
where
(6)
for i, j = 1, 2, …, n. Then, there exist
,
and
such that
(7)
MCGNN (1) is referred to as the drive system, this is the corresponding response system:
(8)
where Ri(t) is the appropriate controller; the initial value of MCGNN (8) is ϕ(s) = (ϕ1(s), ϕ2(s), …, ϕn(s))T, s ≤ 0; bij(ηi(t)), cij(ηi(t)) and dij(ηi(t)) are defined as:
(9)
for i, j = 1, 2, …, n.
In this paper, we design such a feedback controller:
(10)
where pi > 0 and qi > 0 are control gains.
Similarly, it can be derived that
(11)
where yi(t) = Φi(ηi(t)),
,
,
and
(12)
for i, j = 1, 2, …, n.
Let ei(t) = yi(t) − xi(t), i = 1, 2, …, n, then we have
(13)
with initial value ψ(s) = (ψ1(s), ψ2(s), …, ψn(s))T, s ≤ 0, where ψi(s) = Φi(ϕi(s)) − Φi(φi(s)), i = 1, 2, …, n.
Lemma 1 [49]. (Chain Rule) If V(⋅) : Rn → R is C−regular and x(t) ∈ Rn is absolutely continuous on any compact subinterval of [0, +∞), then V(x(t)): [0, +∞) → R is differentiable for a.e.t ∈ [0, +∞) and
where ∂V(x(t)) is the Clarke generalized gradient.
Definition 1. MCGNN (8) is said to be synchronized with MCGNN (1) in finite time, if there exists a constant t* > 0 such that and ei(t) ≡ 0 for t ≥ t*, i = 1, 2, …, n, where t* is called the settling time.
Remark 1. and ei(t) ≡ 0 for t ≥ t* mean that
and xi(t) ≡ yi(t) for t ≥ t*, that is to say,
and Φi(ξi(t)) ≡ Φi(ηi(t)) for t ≥ t*. Since Φi(⋅) is strictly monotone increasing, we know
and ei(t) ≡ 0 for t ≥ t* are also equivalent to
and ξi(t) ≡ ηi(t) for t ≥ t*.
Main results
In this section, we will derive some sufficient conditions that can guarantee the finite time synchronization of MCGNNs (1) and (8).
Theorem 1. Let assumptions A1-A6 hold. If control gains pi and qi satisfy
(14)
MCGNN (8) will be synchronized with MCGNN (1) in finite time under the controller (10).
Proof. We design such a Lyapunov function:
(15)
where
(16)
By Lemma 1, the derivative of V1(t) can be calculated as:
(17)
Based on assumptions A2 and A3, it can be obtained that
(18)
where θ1 is between
and
, θ2 is between yi(t) and xi(t).
Based on assumptions A2 and A4, it follows that
(19)
Similarly, we have
(20)
and
(21)
Based on assumption A5, it follows that
(22)
where γi = 0 if ei(t) = 0, otherwise γi = 1. Similarly, we get
(23)
and
(24)
Furthermore, since is strictly monotone increasing, we know that
. Then we have
(26)
Calculating the derivatives of V2(t) and V3(t), we get that
(27)
and
(28)
where assumptions A1 and A6 have been used.
If the conditions in Theorem 1 are satisfied, we have
(30)
where
By using the same analysis methods as those in [44], we can prove there exists a constant t* > 0 such that
(31)
where e(t) = (e1(t), e2(t), …, en(t))T and
According to Definition 1, MCGNNs (1) and (8) achieve synchronization in finite time. The proof is completed.
Remark 2. In Theorem 1, since the distributed delays in MCGNNs (1) and (8) are unbounded, it is difficult to estimate the settling time t*.
If the delay kernels satisfy
(32)
where βij > 0 are constants, i, j = 1, 2, …, n, MCGNN (1) can be written as
(33)
This is the corresponding response system:
(34)
In fact, MCGNNs (33) and (34) can also achieve finite time synchronization under the controller (10), what is more, the settling time t* can be estimated.
Corollary 1. Let assumptions A1-A5 hold. If control gains pi and qi satisfy
(35)
MCGNN (34) will be synchronized with MCGNN (33) in finite time under the controller (10). Moreover, the settling time
where
Proof. Consider such a Lyapunov function:
(36)
where
(37)
Referring to the proofs of Theorem 1 and Ref. [44], we can give the remaining proof of Corollary 1, which is omitted here.
Remark 3. In MCGNN (1), if is replaced by
, where 0 ≤ ρij(t) ≤ ρij, we can get a new MCGNN. Similarly to Corollary 1, we can prove that MCGNNs with this kind of distributed time-varying delays can achieve finite time synchronization under the controller (10), and the settling time t* can also be estimated.
Remark 4. It has been proved that controller (10) can synchronize MCGNNs effectively. Controller (10) consists of two parts: linear part −pi(ηi(t) − ξi(t)) and nonlinear part −qisign(ηi(t) − ξi(t)). In the proofs of Theorem 1 and Corollary 1, the nonlinear part of the controller is used to deal with the parameter mismatches of the drive-response MCGNNs, while the linear part of the controller plays a key role in driving the response MCGNN to synchronize with the drive MCGNN.
Remark 5. In [43], the authors also investigated the finite time synchronization of MCGNNs. However, the finite-time synchronization analysis methods they utilized were traditional ones [50], that is, they should prove , α > 0, 0 < η < 1, or
, α > 0, θ > 0, 0 < η < 1, where V(t) is the Lyapunov function. In this paper, we utilize some novel finite-time synchronization analysis methods [44]. First, we prove that
, where ε > 0; γi = 0 if ei(t) = 0, otherwise γi = 1. Then we use the strict mathematic analysis to derive the results. Moreover, though the delays considered in [43] were only discrete delays, the controller used in [43] was very complicated, i.e.
. In this paper, we consider MCGNN model with mixed delays, however, the controller that we use is very simple, i.e. Ri(t) = −pi(ηi(t) − ξi(t)) − qisign(ηi(t) − ξi(t)).
Remark 6. In MCGNN (1), if the memristive connection weights bij(ξi(t)) = bij, cij(ξi(t)) = cij and dij(ξi(t)) = 0, MCGNN (1) will reduce into the Cohen-Grossberg neural network model studied in [39, 40]. Therefore, the theoretical results of this paper can be applicable to the Cohen-Grossberg neural networks in [39, 40], while the opposite is probably not true. In this sense, the obtained results of this paper are less conservative.
Numerical simulations
In this section, numerical simulations are given to validate the obtained results in this paper.
Example 1. Consider the following MCGNN:
(38)
where a1(v) = 1.8v, a2(v) = 1.6v, τ11(t) = 1 − 0.2sint, τ12(t) = 0.9 − 0.1cost, τ21(t) = 0.5sint, τ22(t) = 0.5cost, I1 = −0.02, I2 = −0.12,
, fj(v) = tanh(v), Kij(t) = e−0.5t, i, j = 1, 2, and Υ1 = Υ2 = 1,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
.
The initial value of MCGNN (38) is φ(t) = (−0.2, 1.2)T for t ∈ [−5, 0], and φ(t) = (0, 0)T for t ∈ (−∞, −5). The transient behaviour of MCGNN (38) is showed in Fig 2.
This is the corresponding response system:
(39)
The initial value of MCGNN (39) is ϕ(t) = (0.4, 0.6)T for t ∈ [−5, 0], and ϕ(t) = (0, 0)T for t ∈ (−∞, −5). The transient behaviour of MCGNN (39) without control inputs is showed in Fig 3.
The synchronization errors between MCGNNs (38) and (39) are defined as zi(t) = ηi(t) − ξi(t), i = 1, 2. The evolutions of the synchronization errors between MCGNNs (38) and (39) without control inputs are showed in Fig 4.
Obviously, τ11 = 1.2, τ12 = 1, τ21 = 0.5, τ22 = 0.5, σ11 = 0.2, σ12 = 0.1, σ21 = 0.5, σ22 = 0.5, a1 = 1.8, a2 = 1.6, ,
, li = 1, Mi = 1, Kij = 2, i, j = 1, 2, so assumptions A1-A6 hold. According to Theorem 1, if we choose p1 = 7, p2 = 6.6, q1 = 5.4 and q2 = 6.2, MCGNN (39) will be synchronized with MCGNN (38) in finite time under the controller (10). Fig 5 shows the evolutions of the synchronization errors between MCGNNs (38) and (39) under the controller (10).
Example 2. Consider the following MCGNN:
(40)
where a1(v) = 1.61v + sin(v), a2(v) = 1.45v + sin(v), I1 = I2 = 0,
,
, Υ1 = 0.3, Υ2 = 1,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
,
, and fj(⋅), τij(t), i, j = 1, 2, are the same as those in Example 1. The initial value of MCGNN (40) is φ(t) = (−0.2, 1.2)T, t ∈ [−2, 0]. The transient behaviour of MCGNN (40) is showed in Fig 6.
This is the corresponding response system:
(41)
The initial value of MCGNN (41) is ϕ(t) = (0.4, 0.6)T, t ∈ [−2, 0]. The transient behaviour of MCGNN (41) without control inputs is showed in Fig 7. The evolutions of the synchronization errors between MCGNNs (40) and (41) without control inputs are showed in Fig 8.
According to Corollary 1, if we choose p1 = 48.3, p2 = 45.6, q1 = 20 and q2 = 10, MCGNN (41) will be synchronized with MCGNN (40) in finite time under the controller (10), and the settling time t* can be estimated as 7.143. Fig 9 shows that MCGNNs (40) and (41) realize finite time synchronization within t*.
Conclusion
This paper studies the finite time synchronization problem of MCGNNs with mixed delays. By utilizing some novel and effective analysis techniques, several sufficient conditions that can guarantee the finite time synchronization of MCGNNs with mixed delays are derived. The feedback controllers that we design are very simple, but they can solve the parameter mismatch problem of the drive-response MCGNNs perfectly. However, the conservativeness of the theoretical analysis probably makes the control gains of our feedback controllers much larger than those needed in the engineering applications. On the other hand, it is costly and impractical to control a network by applying controllers to all the nodes. Since adaptive pinning controller can avoid the high control gains effectively and reduce the number of the controlled nodes, our future work will focus on the synchronization control of MCGNNs via the adaptive pinning control. Numerical simulations are given to verify the obtained theoretical results.
Acknowledgments
The work is supported by the National Key Research and Development Program (Grant nos. 2016YFB0800602 and 2016YFB0800604), and the National Natural Science Foundation of China (Grant Nos. 61573067, 61771071 and 61472045).
References
- 1. Chua LO. Memristor-the missing circuit element. IEEE Transactions on Circuit Theory. 1971; 18: 507–519.
- 2. Struko DB, Snider GS, Stewart GR, Williams RS. The missing memristor found. Nature. 2008; 453: 80–83.
- 3. Wang WP, Li LX, Peng HP, Xiao JH, Yang YX. Synchronization control of memristor-based recurrent neural networks with perturbations. Neural Networks. 2014; 53: 8–14. pmid:24524891
- 4. Jo SH, Chang T, Ebong I, Bhadviya BB, Mazumder P, Lu W. Nanoscale memristor device as synapse in neuromorphic systems. Nano Letters. 2010; 10: 1297–1301. pmid:20192230
- 5. Zhang H, Wang XY, Lin XH. Topology identification and module-phase synchronization of neural network with time delay. IEEE Transactions on Systems Man and Cybernetics Systems. 2017; 47(6): 885–892.
- 6. Tang Q, Wang XY. Backstepping generalized synchronization for neural network with delays based on tracing control method. Neural Computing and Applications. 2014; 24: 775–778.
- 7.
Jiao B, Wu XQ. The 3-cycle weighted spectral distribution in evolving community-based networks. Chaos. 2017; https://doi.org/10.1063/1.4978024
- 8. Wen SP, Zeng ZG, Huang TW, Meng QG, Yao W. Lag Synchronization of Switched Neural Networks via Neural Activation Function and Applications in Image Encryption. IEEE Transactions on Neural Networks and Learning Systems. 2015; 26(7): 1493–1502. pmid:25594985
- 9. Milanović V, Zaghloul ME. Synchronization of chaotic neural networks and applications to communications. International Journal of Bifurcation and Chaos. 1996; 6: 2571–2585.
- 10. Tan Z, Ali MK. Associative memory using synchronization in a chaotic neural network. International Journal of Modern Physics C. 2001; 12(1): 19–29.
- 11. Itoh M, Chua LO. Memristor oscillators. International Journal of Bifurcation and Chaos. 2008; 18: 3183–3206.
- 12. Wang XY, Wang MJ. A hyperchaos generated from lorenz system. Physica A Statistical Mechanics and Its Applications. 2008; 387(14): 3751–3758.
- 13. Wang CN, Chu RT, Ma J. Controlling a chaotic resonator by means of dynamic track control. Complexity. 2015; 21(1): 370–378.
- 14. Han XM, Wu HQ, Fang BL. Adaptive exponential synchronization of memristive neural networks with mixed time-varying delays. Neurocomputing. 2016; 201: 40–50.
- 15. Zhang W, Li CD, Huang TW, Huang JJ. Stability and synchronization of memristor-based coupling neural networks with time-varying delays via intermittent control. Neurocomputing. 2016; 173: 1066–1072.
- 16.
Chen C, Li LX, Peng HP, Yang YX, Li T. Synchronization control of coupled memristor-based neural networks with mixed delays and stochastic perturbations. Neural Processing Letters. 2017; https://doi.org/10.1007/s11063-017-9675-6
- 17. Xiao JY, Zhong SM, Li YT, Xu F. Finite-time Mittag-Leffler synchronization of fractional-order memristive BAM neural networks with time delays. Neurocomputing. 2016; 219: 431–439.
- 18. Yang XS, Ho DWC. Synchronization of delayed memristive neural networks: robust analysis approach. IEEE Transactions on Cybernetics. 2015; 46(12): 3377–3387.
- 19. Yang XS, Ho DWC, Lu JQ, Song Q. Finite-time cluster synchronization of t-s fuzzy complex networks with discontinuous subsystems and random coupling delays. IEEE Transactions on Fuzzy Systems. 2015; 23(6): 2302–2316.
- 20. Yang XS. Can neural networks with arbitrary delays be finite-timely synchronized? Neurocomputing. 2014; 143(16): 275–281.
- 21. Yang XS, Song Q, Liang JL, He B. Finite-time synchronization of coupled discontinuous neural networks with mixed delays and nonidentical perturbations. Journal of the Franklin Institute. 2015; 352(10): 1–30.
- 22. Wang XY, Song JM. Synchronization of the fractional order hyperchaos lorenz systems with activation feedback control. Communications in Nonlinear Science and Numerical Simulation. 2009; 14(8): 3351–3357.
- 23. Li LX, Li WW, Kurths J, Luo Q, Yang YX, Li SD. Pinning adaptive synchronization of a class of uncertain complex dynamical networks with multi-link against network deterioration. Chaos Solitons Fractals. 2015; 72: 20–34.
- 24. Mei GF, Wu XQ, Ning D, Lu JA. Finite-time stabilization of complex dynamical networks via optimal control. Complexity. 2016; 21(1): 417–425.
- 25. Wang XY, He Y. Projective synchronization of fractional order chaotic system based on linear separation. Physics Letters A. 2008; 372(4): 435–441.
- 26. Wang XY, Wang MJ. Dynamic analysis of the fractional-order Liu system and its synchronization. Chaos. 2007; 17(3): 304–311.
- 27. Yang XS, Huang CX, Zhu QX. Synchronization of switched neural networks with mixed delays via impulsive control. Chaos Solitons Fractals. 2011; 44(10):817–826.
- 28. Lin D, Wang XY, Nian FZ, Zhang YL. Dynamic fuzzy neural networks modeling and adaptive backstepping tracking control of uncertain chaotic systems. Neurocomputing. 2010; 73(16): 2873–2881.
- 29. Lin D, Wang XY. Observer-based decentralized fuzzy neural sliding mode control for interconnected unknown chaotic systems via network structure adaptation. Fuzzy Sets and Systems. 2010; 161(15): 2066–2080.
- 30. Cohen MA, Grossberg S. Absolute stability of global pattern formation and parallel memory storage by competitive neural networks. IEEE Transactions on Systems Man and Cybernetics. 1983; 13: 815–826.
- 31. Zhu QX, Li XD. Exponential and almost sure exponential stability of stochastic fuzzy delayed Cohen-Grossberg neural networks. Fuzzy Sets Systems. 2012; 203: 74–94.
- 32. Tojtovska B, Jankovic S. On a general decay stability of stochastic Cohen-Grossberg neural networks with time-varying delays. Applied Mathematics and Computation. 2012; 219: 2289–2302.
- 33. Zheng CD, Shan QH, Zhang HG, Wang ZS. On stabilization of stochastic Cohen-Grossberg neural networks with mode-dependent mixed time-delays and Markovian switching. IEEE Transactions on Neural Networks and Learning Systems. 2013; 24: 800–811. pmid:24808429
- 34. Wang ZD, Liu YR, Li MZ, Liu XH. Stability analysis for stochastic cohen-grossberg neural networks with mixed time delays. IEEE Transactions on Neural Networks. 2006; 17(3): 814–820. pmid:16722186
- 35. Zhu QX, Cao JD. Adaptive synchronization of chaotic Cohen-Grossberg neural networks with mixed time delays. Nonlinear Dynamics. 2010; 61: 517–534.
- 36. Yu J, Hu C, Jiang HJ, Teng ZD. Exponential synchronization of Cohen-Grossberg neural networks via periodically intermittent control. Neurocomputing. 2011; 74: 1776–1782.
- 37. Liu QM, Zhang SH. Adaptive lag synchronization of chaotic Cohen-Grossberg neural networks with discrete delays. Chaos. 2012; 22: 261–271.
- 38. Zhu QX, Cao JD. pth moment exponential synchronization for stochastic delayed Cohen-Grossberg neural networks with Markovian switching. Nonlinear Dynamics. 2012; 67: 829–845.
- 39. Wan Y, Cao JD, Wen GH, Yu WW. Robust fixed-time synchronization of delayed Cohen-Grossberg neural networks. Neural Networks. 2016; 73: 86–94. pmid:26575975
- 40. Hu C, Yu J, Jiang HJ. Finite-time synchronization of delayed neural networks with cohen-grossberg type based on delayed feedback control. Neurocomputing. 2014; 143(16): 90–96.
- 41. Yang XS, Cao JD, Yu WW. Exponential synchronization of memristive cohen-grossberg neural networks with mixed delays. Cognitive Neurodynamics. 2014; 8(3): 239–249. pmid:24808932
- 42. Abdurahman A, Jiang HJ, Rahman K. Function projective synchronization of memristor-based cohen-grossberg neural networks with time-varying delays. Cognitive Neurodynamics. 2015; 9(6): 1–11.
- 43. Liu M, Jiang HJ, Hu C. Finite-time synchronization of memristor-based cohen-grossberg neural networks with time-varying delays. Neurocomputing. 2016; 194: 1–9.
- 44. Chen C, Li LX, Peng HP, Yang YX, Li T. Finite-time synchronization of memristor-based neural networks with mixed delays. Neurocomputing. 2017; 235: 83–89.
- 45. Wu AL, Zeng ZG. Dynamic behaviors of memristor-based recurrent neural networks with time-varying delays. Neural Networks. 2012; 36: 1–10. pmid:23037770
- 46. Wu AL, Zeng ZG. Anti-synchronization control of a class of memristive recurrent neural networks. Communications in Nonlinear Science and Numerical Simulation. 2013; 18: 373–385.
- 47.
Filippov AF. Differential Equations with Discontinuous Right-hand Side. Boston: Mathematics and Its Applications (Soviet Series), Kluwer Academic; 1988.
- 48.
Aubin JP, Cellina A. Differential Inclusions. Berlin: Springer-Verlag; 1984.
- 49. Clarke FH. Nonsmooth analysis and optimization. In Proceedings of the international congress of mathematicians. 1984; 5: 847–853.
- 50. Tang Y. Terminal sliding mode control for rigid robots. Automatica. 1998; 34(1): 51–56.