Further results on dissipativity analysis of neural networks with time-varying delay and randomly occurring uncertainties

In this paper, the problem of robust dissipativity is investigated for neural networks with both time-varying delay and randomly occurring uncertainties. The randomly occurring uncertainties are assumed to obey mutually uncorrelated Bernoulli-distributed white noise sequences. By constructing a new Lyapunov-Krasovskii functional, some improved delay-dependent dissipativity conditions are derived based on two integral inequalities, which are formulated in terms of linear matrix inequality. Furthermore, some information of activation function ignored in previous works has been taken into account in the resulting condition. The effectiveness and the improvement of the proposed approach are demonstrated by two illustrating numerical examples.