一种不定控制系统的强健神经网络控制器

探讨不定控制系统的强健神经网络控制器．这是一个由遗传运算法则确定的最佳隐藏神经数的前馈多层神经网络．它借助状态空间控制器和编入必要的最佳程序，进行非线性处理，以确保性能的强健性．混合的遗传运算法则和后繁殖神经培养，具有最佳网络结构和神经染色体接合权的双重目标，把最佳网络结构和神经权协调地聚合，从而得到与性能控制目标基本一致．

A ROBUST NEURAL NETWORK CONTROLLER FOR UNCERTAIN CONTROL SYSTEMS

A Robust neural network controller for uncertain control systems has been developed in this paper. This neural controller is a feed-forward multiplayer neural network with its optimal number of hidden neurons determined by a genetic algorithm. It is based upon the state-space controller, and incorporates necessary optimization procedures to handle significant nonlinearities and to guarantee performance neural training with the dual objectives of optimizing the network structure and neural weights which yield a low error in accordance to the performance control objective.