錢學(xué)明
(1.無錫科技職業(yè)學(xué)院物聯(lián)網(wǎng)技術(shù)學(xué)院,江蘇無錫 214028;2.江南大學(xué)物聯(lián)網(wǎng)工程學(xué)院,江蘇無錫 214122)
離散時間的混合時滯耦合神經(jīng)網(wǎng)絡(luò)的魯棒指數(shù)同步
錢學(xué)明1,2
(1.無錫科技職業(yè)學(xué)院物聯(lián)網(wǎng)技術(shù)學(xué)院,江蘇無錫 214028;2.江南大學(xué)物聯(lián)網(wǎng)工程學(xué)院,江蘇無錫 214122)
討論了一類離散時間的時滯耦合神經(jīng)網(wǎng)絡(luò)的同步問題.在參數(shù)不確定的離散時間耦合神經(jīng)網(wǎng)絡(luò)中,考慮了變時滯和有限分布時滯.同時,細胞激活函數(shù)假設(shè)為較Lipschitz條件更為一般的扇形非線性函數(shù),該函數(shù)可以既不可微又不嚴格單調(diào).通過構(gòu)造Lyapunov-Krasovskii泛函,運用線性矩陣不等式(LM I)技術(shù),并結(jié)合Kronecker積來獲得耦合神經(jīng)網(wǎng)絡(luò)魯棒全局指數(shù)同步的充分性判據(jù),并且所獲得的判據(jù)依賴于時滯.最后,對一個實例進行仿真,說明結(jié)論的有效性.
耦合神經(jīng)網(wǎng)絡(luò)系統(tǒng);離散時間系統(tǒng);混合時滯;同步;魯棒
自從Pecora和Carroll引入了一種能使兩個初始條件不同的混沌系統(tǒng)達到同步的方法后,混沌同步已逐漸成為非線性領(lǐng)域中最受關(guān)注的研究之一.近年來,耦合系統(tǒng)的同步由于其在基因網(wǎng)絡(luò)、圖像處理、保密通信等領(lǐng)域的廣泛應(yīng)用,引起了越來越多專家學(xué)者的興趣.人們相繼研究了很多連續(xù)時間耦合系統(tǒng)的同步問題[1-3].然而,在對連續(xù)時間神經(jīng)網(wǎng)絡(luò)進行實驗、仿真、計算時,通常需要對其離散化.但正如文[4]中提出的,即使在一個小的取樣周期下,離散化也不能保證系統(tǒng)連續(xù)部分的性質(zhì).因此,研究離散時間的神經(jīng)網(wǎng)絡(luò)至關(guān)重要.文[5]分析了一類離散時間耦合系統(tǒng)的同步現(xiàn)象及其動力學(xué)過程.
由于在現(xiàn)實世界中時滯現(xiàn)象不可避免,文[6-7]相繼研究了含變時滯的離散時間神經(jīng)網(wǎng)絡(luò),并給出了依賴時滯的同步判據(jù).其后,離散時間的時滯耦合神經(jīng)網(wǎng)絡(luò)[8-9]也得到了深入的研究.2008年,文[10]在研究離散時間復(fù)雜網(wǎng)絡(luò)時首次引入了無窮分布時滯.文[11]進一步考慮了含有無窮分布時滯的耦合神經(jīng)網(wǎng)絡(luò),并給出了LM I形式的全局漸近同步判據(jù).然而,在離散時間的時滯耦合神經(jīng)網(wǎng)絡(luò)中引入有限分布時滯,并考慮系統(tǒng)參數(shù)不確定的情形下,系統(tǒng)魯棒指數(shù)同步的研究至今還鮮見于相關(guān)文獻,是一個值得討論的課題.本文擬研究一類離散時間的混合時滯耦合神經(jīng)網(wǎng)絡(luò)的魯棒指數(shù)同步性.本文中表示維Euclid空間,表示所有的實矩陣構(gòu)成的集合.意味著是半正定的,意味著是正定的.表示具有適當維數(shù)的單位矩陣.表示中的Euclid向量范數(shù).分別表示矩陣的最大特征值和最小特征值.一個m n×矩陣X和一個p q×矩陣Y的Kronecker積定義為一個矩陣,記為.星號*在矩陣中用來表示與之對稱的項.
注:若在耦合神經(jīng)網(wǎng)絡(luò)(1)或(30)中將有限分布時滯改為無窮分布時滯,則沿用定理1的證明方法,即可得類似的結(jié)論.
從而,由定理1可知離散時間的混合時滯耦合神經(jīng)網(wǎng)絡(luò)(1)是魯棒全局指數(shù)同步的.
本文研究了一類離散時間的時滯耦合神經(jīng)網(wǎng)絡(luò)的同步問題.在模型中,考慮了變時滯和有限離散分布時滯以及系統(tǒng)參數(shù)的不確定性等方面的影響.此外,細胞激活函數(shù)描述為扇形非線性函數(shù),可以既不可微又不嚴格單調(diào),甚至是無界的.該條件比目前廣泛使用的Lipschitz條件更為一般,可以降低結(jié)論的保守性.本文利用Lyapunov函數(shù)方法結(jié)合Kronecker積來獲得離散時間的時滯耦合神經(jīng)網(wǎng)絡(luò)的魯棒全局同步的充分性判據(jù),而且所獲得的LM I形式的判據(jù)依賴于變時滯和有限分布時滯的上界和下界.該判據(jù)易于用MATLAB的LM I工具箱進行有效的驗證.最后,數(shù)值仿真的實例可以說明我們所得到的同步判據(jù)的有效性和可應(yīng)用性.
[1] Wu C W, Chua L O. Synchronization in an array of linearly coupled dynam ical system s [J]. IEEE Trans. On Circuits and Systems, 1995, 42 (8): 430-447.
[2] Cao J D, Li P, Wang W W. Global synchronization in arrays of delayed neural networks with constant and delayed coupling [J]. Physic Letters A, 2006, 353(4), 318-325.
[3] Liu Y R, Wang Z D, Liu X H. On synchronization of coup led neural netw orks w ith discrete and unbounded distributed delays [J]. International Journal of Computer Mathematics, 2008, 85(8): 1299 -1313.
[4] Wu C W. Synchronization in arrays of coupled nonlinear systems w ith delay and nonreciprocal time-varying coupling [J]. IEEE Trans. Circuits Syst.-II, 2005, 52(5): 282-286.
[5] Lu W L, Chen T P. Synchronization analysis of linearly coupled networks of discrete time systems [J]. Physica D, 2004, 198(1): 148-168.
[6] Liang J L, Cao J D, Daniel W C H. Discrete-time bidirectional associative memory neural networks with variable delays [J]. Physics Letters A, 2005, 335(2): 226-234.
[7] Song Q K, Wang Z D. A delay-dependent LMI approach to dynamics analysis of discrete-time recurrent neural networks with time-varying delays [J]. Physics Letters A, 2007, 368(1): 134-145.
[8] Li T, Song A G, Fei S M. Synchronization control for arrays of coupled discrete-time delayed cohen-grossberg neural networks [J]. Neurocomputing, 2010, 74(1): 197-204.
[9] Liang J L, Wang Z D, Liu Y R, et al. Robust synchronization of an array of coupled stochastic discrete-time delayed neural networks [J]. IEEE Trans. On Neural Networks, 2008, 19(11): 1910-1921.
[10] Liu Y R, Wang Z D, Liang J L, et al. Synchronization and state estimation for discrete-time complex networks with distributed delays [J]. IEEE Trans. On Systems, M an and Cybernetics-Part B: Cybernetics, 2008, 38(5): 1314-1325.
[11] Wang H W, Song Q K. Synchronization for an array of coupled stochastic discrete-time neural networks with mixed delays [J]. Neurocomputing, 2011, 74(10): 1572-1584.
Robust Exponential Synchronization in an Array of Discrete-Time Coupled Neural Networks w ith M ixed Time Delays
QIAN Xuem ing1,2
(1. School of Internet of Things, Wuxi Professional College of Science and Technology, Wuxi, China 214028; 2. School of Internet of Things, Jiangnan University, Wuxi, China 214122)
This paper addresses the analysis problem of synchronization for a class of discrete-time coupled neural networks w ith time-varying and distributed delays. The neural networks are subject to parameter uncertainty. Furthermore, the description of the activation functions is a more general sector nonlinear function than the recently commonly-used Lipschitz conditions, which are assumed to be neither differentiable nor strictly monotonic. By referring to Lyapunov functional method and K ronecker product technique, some sufficient conditions depending on delay are derived for robust exponential synchronization of such systems. Finally, a simulation example is presented to show the usefulness of the derived LM I-based synchronization scheme.
Coupled Neural Networks Systems; Discrete-time Systems; M ixed Time Delays; Synchronization; Robust
O175;TP183
:A
:1674-3563(2014)03-0001-11
10.3875/j.issn.1674-3563.2014.03.001 本文的PDF文件可以從xuebao.wzu.edu.cn獲得
(編輯:封毅)
2013-09-29
江蘇省自然科學(xué)基金(BK2010313)
錢學(xué)明(1981- ),男,江蘇無錫人,博士研究生,講師,研究方向:復(fù)雜系統(tǒng)的控制理論