張 韜,蘇亞坤,朱 進(jìn)
(渤海大學(xué) 數(shù)理學(xué)院,遼寧錦州 121000)
Cohen-Grossberg神經(jīng)網(wǎng)絡(luò)是由Cohen和Crossberg于1983年提出的[1],被廣泛地應(yīng)用于模式識別、記憶與信號處理、圖象處理與計(jì)算技術(shù)等領(lǐng)域。然而,在實(shí)際應(yīng)用中時(shí)滯、脈沖是不可避免的,且時(shí)滯、脈沖對神經(jīng)網(wǎng)絡(luò)的穩(wěn)定性有著巨大的影響[2-8],因此有關(guān)時(shí)滯脈沖Cohen-Crossberg神經(jīng)網(wǎng)絡(luò)的研究[9-19]已逐漸引起人們的關(guān)注,研究脈沖型時(shí)滯神經(jīng)網(wǎng)絡(luò)具有極其重要的意義。
神經(jīng)網(wǎng)絡(luò)模型如下:
初始條件x(t0+s)=φ(s),0≤τij(t)<τij(t)≤η<1,其中:x(t)=(x1(t),x2(t),…,xn(t))表示神經(jīng)元狀態(tài)向量;ai(·)表示放大函數(shù);bi(·)表示適當(dāng)?shù)男袨楹瘮?shù);fj,hj為神經(jīng)元的激勵(lì)函數(shù);C=(cij)n×n,D=(dij)n×n,W=(wij)n×n分別表示連接權(quán)矩陣、時(shí)滯連接權(quán)矩陣和分布時(shí)滯連接權(quán)矩陣。固定時(shí)刻tk滿足t1<t2<t3<…,且在 tk時(shí)刻,Δ x(t )Rn」表示在tk時(shí)刻的狀態(tài)變化,對所有的k∈N,Ik(0)=0。
要求神經(jīng)網(wǎng)絡(luò)模型滿足以下假設(shè):
1)存在正常數(shù) Lj,Hj,j=1,2,…,n 使得
4)?σk≥0,k∈N,有
5)?μ >1,有 μτ≤inf{tk-tk-1};
6)max{ θk}≤M < e2λμτ,M 是常數(shù),θk=1+(2σk+);
7)延遲核函數(shù) Kij,i,j=1,2…n是定義在[0,∞)上的實(shí)值非負(fù)函數(shù),滿足,其中λ是正常數(shù)。
定理 在假設(shè)1)~7)下,如果?λ>0,正對角矩陣Q=diag(q1,…,qn),使得
其中
那么模型(1)的零解是全局指數(shù)穩(wěn)定的。
證明 構(gòu)造如下Lyapunov-Krasovskii泛函:
當(dāng) t≠tk時(shí)
利用條件1)~3)和2ab≤a2+b2得
則
由V'<0 知函數(shù) V(t,x(t))是單調(diào)遞減的,有 V(t,x(t))≤V(t0,x(t0)),
又因?yàn)?/p>
當(dāng)t=tk時(shí),根據(jù)假設(shè)4)~6)和指數(shù)穩(wěn)定定義,有
由模型的任意解x(t,t0,x0)可得
由于μτ≤inf{tk-tk-1},μτ≤t1-t0,μτ≤t2-t1,…,μτ≤tk-1-tk-2,求和得 (k-1)μτ≤t1-t0+t2-t1+…
考慮下面的系統(tǒng)
其中 a1(x1(t))=3+sinx1(t),a2(x2(t))=4+cosx1(t)。
[1]Cohen M A,Grossberg S.Absolute stability of global pattern formation and parallel memory storage by competitive neural networks[J].IEEE Trans Syst Man Cybern,1983,13(5):815-826.
[2]Liao X,Li C.Global attractivity of Cohen-Grossberg model with finite and infinite delays[J].Math Anal Appl,2006,315(1):244-262.
[3]Huang T,Chan A,Huang Y,et al.Stability of Cohen-Grossberg neural networks with time varying delays[J].Neural Networks,2007,20(8):868-873.
[4]Li T,F(xiàn)ei S.Stability analysis of Cohen-Grosserg neural networks with timevarying and distributed delays[J].Neurocomputing,2008,71:1069-1081.
[5]Li K.Stability analysis for impulsive Cohen-Grossberg neural networks with time-varying delays and distributed delay[J].Nonlinear Anal Real World Appl,2009,10(5):2784-2798.
[6]Zheng C D,Shan Q H,Wang Z.Novel stability criteria of Cohen-Grossberg neural networks with time-varying delays[J].Int J Circuit Theory Appl,2012,40(3):221-235.
[7]Huang C X,Huang L H,He Y G.Mean square exponential stability of stochastic Cohen-Grossberg neural networks with unbounded distributed delays[J].Discrete Dynamics in Nature and Society,2010(20):1-15.
[8]Wan L,Zhou Q.Exponential stability of stochastic reaction diffusion Cohen-Grossberg neural networks with delays[J].Applied Mathematics and Computation,2008,206(2):818-824.
[9]Chen Y.Global Asymptotic Stability of Delayed Cohen-Grossberg Neural Networks[J].IEEE Transactions on Circuit and Systems-I:Fundamental Theory and Applications,2006,53(2):351-357.
[10]Yang Z C,Xu D Y.Impulsive effects on stability of Cohen-Grossberg neural nettworks with variable delays[J].Applied Mathematics and Cumputation,2006,177(1):63-78.
[11]Bai C Z.Stability analysis of Cohen-Grossberg BAM neural networks with delays and impulses[J].Chaos,Solitons and Fractals,2008,35(2):263-267.
[12]Ping Z W,Lu J G.Global exponential stability of impulsive Cohen-Grossberg neural networks with continuously distributed delays[J].Chaos,Solitons and Fractals,2009,41(1):164-174.
[13]Luo W P,Zhong S M,Yang J.Global exponential stability of impulsive Cohen-Grossberg neural networks with delays[J].Chaos,Solitons and Fractals,2009,42(2):1084-1091.
[14]Li K L.Stability analysis for impulsive Cohen Grossberg neural networks with time varying delays and distributed delays[J].Nonlinear Analysis:Real World Applications,2009,10(5):2784-2798.
[15]Lou X Y,Cui B T.Global exponential stability analysis of delayed Cohen-Grossberg neural networks with distributed delay[J].International Journal of Systems Science,2007,38(7):601-609.
[16]Song Q K,Cao J D.Robust Stability in Cohen Grossberg Neural Network with both Time Varying and Distributed Delays[J].Neural Process Lett,2008,27(2):179-196.
[17]Wang B X,Jian J G,Jiang M H.Stability in Lagrange sense for Cohen-Grossberg neural networks with time-varying delays and finite distributed delays[J].Nonlinear AnaIysis:Hybrid Systems,2010,4(1):65-78.
[18]Wu W,Cui B T,Lou X Y.Global exponential stability of Cohen-Grossberg neural networks with distributed delays[J].Mathematical and Computer Modelling,2008,47(9):868-873.
[19]Lu K,Xu D,Yang Z.Global attraction and stability for Cohen-Grossberg neural networks with delays[J].Neural Networks,2006,19(10):1538-1549.