Synchronization of neural networks based on parameter identification and via output or state coupling (Q955063)

Let \[ \dot x(t)=- C_1 x(t)+ A_1 f(x(t))+ B_1 f(t-\tau(t))+ J\tag{i} \] be a controlling neural network and \[ \dot y(t)= -C_2 y(t)+ A_2 f(y(t)+ B_2 f(t-\tau(t))+ J+ U(t)\tag{ii} \] a controllable neural network where \(U(t)\) is a control vector, \(x,y\in \mathbb{R}^n\), \(0\leq t<\infty\). The authors design feedback and adaptive state coupling controls and adaptive estimates for \(\dot A_2\), \(\dot B_2\), \(\dot C_2\) such that the system (ii) is synchronous to system (i). The method used is to construct Lyapunov functionals \(V(t)\) such that \(\dot V(t)= 0\) if and only if \(e(t)= 0\) where \(e(t)= x(t)- y(t)\) is the synchronization error. This implies that \(e(t)\to 0\), \(t\to\infty\). Numerical results are presented in which \(n= 2\) and (i) is a chaotic delayed Hopfield neural network. Both a feedback and state coupling control are determined such that the controlled systems are synchronous to the original system.