New stability conditions for neural networks with constant and variable delays (Q2484770)

The inclusion of delay effects in artificial neural networks can eventually produce instability and complex behaviour. This paper gives a new sufficient condition for the asymptotic stability of the equilibrium point of a delayed neural network by constructing an adequate Lyapunov functional. This criterion generalizes some results of \textit{D. Zhou} and \textit{J. Cao} [Appl. Math. Comput. 131, No. 2--3, 487--496 (2002; Zbl 1034.34093)] and \textit{H. Peng} and \textit{A. Schick} [J. Multivariate Anal. 95, No. 2, 385--409 (2005; Zbl 1070.62018)] and is illustrated by two numerical examples.