论文部分内容阅读
研究时变时滞神经网络的鲁棒稳定性和耗散性问题.充分利用积分项的时滞信息和激励函数条件构造一个合适的增广LK泛函;利用自由矩阵积分不等式处理LK泛函的导数,得到一个低保守性的时滞相关稳定判据;将所获得的结论延伸至神经网络的耗散性分析,并推导出一个确保神经网络严格(X,Y,Z)-γ-耗散的充分条件.最后通过3个数值算例验证了所提出方法的可行性和优越性.
To study the robust stability and dissipation of neural networks with time-varying delays, a sufficient augmented LK functional is constructed by fully using the delay information and the incentive function of the integral term. The LK functional is solved by the free matrix integral inequality Derive a low conservative delay-dependent stability criterion, and extend the conclusions obtained to the dissipative analysis of neural networks and derive a new method to ensure the strict (X, Y, Z) -γ-dissipation of neural networks The sufficient conditions of the proposed method are given.Finally, three numerical examples are given to verify the feasibility and superiority of the proposed method.