New Criteria on Stability Analysis of Generalized Neural Networks With Time-Varying Delays
研究了具有时变时滞的广义神经网络的稳定性问题，提出了新的参数相关负定判定条件，放松了李雅普诺夫-克拉索夫斯基泛函的正定约束，得到了更少保守性的稳定性准则，并通过实例验证了方法的可行性和优越性。
The stability problem of generalized neural networks (GNNs) with time-varying delay is investigated in this article. Novel parameter-dependent negative-determination conditions (NDCs) for cubic matrix-valued polynomials are derived by using convex approaches. In comparison with the existing methods, the positive-definite constraint on the constructed Lyapunov–Krasovskii functional (LKF) is relaxed by using the sum of several matrices to maintain positive definiteness instead of this requirement on every single matrix involved in the LKF. Less conservative stability conditions are established by using the constructed LKF and the proposed parameter-dependent NDCs for cubic matrix-valued polynomials. Two examples, including comparative results to existing methods are presented to show the feasibility and superiority of the proposed method.