一种动态大系统实时递阶控制方法

为了克服数值计算方法的困难，给出了一种用于求解动态大系统递阶优化控制问题的神经网络模型，构成该神经网络的局部优化子网络和协调子网络同时工作，给出最优解，具有很高的求解效率，并且该网络具有全局渐近稳定性，便于硬件实现及实际应用。

A New Method for Real-Time Hierarchical Controlof Dynamical Large-Scale Systems

To overcome computational difficulties existing in numerical methods, a neural network model for effectively solving the problems of dynamical hierarchical control of large-scale systems is presented. The local optimization sub -networks and the cordination sub -networks can work simultaneously to provide the optimal solutions to the original problems. The neural network has global stability. lt is suitable for implementation by an integrated-circuit and for real-time applications .