A periodic solution of delayed celluar neural networks

Peiyong Zhu,Shixin Sun
2005-01-01
Neural Network World
Abstract:We mainly prove: Assume that each output function f(i) of DCNN is A bounded on R and satisfies the Lipschitz condition, if Sigma(n)(j=1)(vertical bar a(ij)vertical bar + vertical bar b(ij)vertical bar + vertical bar a(ji)vertical bar + vertical bar b(ji)vertical bar)< < 2c(i)/k and I-i : R+ -> R is a periodic function with period omega for each i, then DCNN k has a unique omega-period solution and all other solutions of DCNN converge exponentially to it, where k = max {k(i) : 1 <= i <= n}, k(i) is a Lipschitz constant of f(i) for i = 1, 2,..., n.
What problem does this paper attempt to address?