基于仿射Bessel-Legendre不等式和不确定转移率的神经网络稳定性

Stability for neural networks based on affine Bessel-Legendre inequality and uncertain transition rates

针对具有时变时滞和不确定转移率的马尔科夫神经网络系统,充分考虑马尔科夫转移率的不确定特性,利用基于松弛变量的有效技术代替传统不等式来约束转移速率中的不确定项,从而减少了决策变量的个数并降低了计算复杂度.通过建立时滞依赖的增广Lyapunov-Krasovskii泛函,并基于仿射Bessel-Legendre(B-L)不等式,给出依赖于时滞和时滞导数上下界的具有较小保守性的神经网络系统稳定条件.最后,通过两个数值例子说明了理论结果的有效性.

For Markovian neural network with time-varying delays and uncertain transition rates,the effective relaxation variable technique instead of the traditional inequality is adopted to restrain the uncertain terms of the transition rates by fully considering the uncertain characteristic of Markovian transition rates,which reduces the number of decision variables and the computational complexity.By applying the delayed-dependent augmented Lyapunov-Krasovskii functional,and affine Bessel-Legendre(B-L)inequality,the less conservative stability conditions that are dependent on upper and lower bounds of delay and delay derivative are proposed.Finally,two numerical examples are presented to illustrate the effectiveness of the theoretical results.