New delay-dependent criterion for the stability of recurrent neural networks with time-varying delay (Q2267138)

A new global asymptotic stability criterion for a class of recurrent neural networks with time-varying delay \[ \dot{u}_i(t) = -a_iu_i(t) +\sum_{j=1}^n \omega_{ij}\bar{g}_j(u_j(t)) + \sum_{j=1}^n\omega_{ij}^1\bar{f}_j(u_j(t-\tau(t))) + U_i \] is presented. Free weighting matrix method is employed to express the relationship between the terms of the Leibniz-Newton formula and a suitable Lyapunov-Krasovskii functional. Obtained results are given in the form of linear matrix inequality (LMI). They depend on the size of time-varying delay, depend on the size of the derivative of time-varying delay. Two numerical examples are presented.