研究等式约束下二次规划问题最优解神经网络模型的稳定性,提出一种变时滞Lagrange神经网络求解方法.利用线性矩阵不等式(LMI)技术,得到两个变时滞神经网络模型全局指数稳定的条件.分析表明,此稳定判据能够适应慢变时滞和快变时滞两种情况,具有适用范围宽、保守性小且易于验证等特点.数值仿真结果验证了所提方法的有效性. In this paper, the stability of the optimal solution neural network model for quadratic programming problems under the constraints of equality constraints is studied, and a Lagrangian neural network with variable delay is proposed. Using the LMI technique, the global stability of two neural networks with time- Exponential stability condition.The analysis shows that the stability criterion can adapt to two cases of slowly varying time lag and fast varying lag.It has the characteristics of wide applicability, small conservativeness and easy to verify.The numerical simulation results verify that the proposed method Effectiveness.