Robust stability criterion for stochastic recurrent neural networks with markovian jumping parameters, mode-dependent delays and multiplicative noise

In this paper, the problem for recurrent neural networks is considered. It is stochastic and contains jumping parameters which are continuous-time Markov process. Delay is mode-dependent and this model is affected by multiplicative noise. Based on the Lyapunov stability theory combined with linear matrix inequalities (LMIs) techniques, we would get some new criteria to guarantee that they are robust stable and their L₂ gains are less than gamma > 0. Introducing into some free weighting matrices would lead to much less conservative results. At last, one numerical example is given to illustrate the effectiveness of the proposed method.