一类非线性随机系统的在线优化算法

An On-line Optimization on Non-linear Stochastic System

讨论一类随机非线性系统的在线优化问题,系统中的非线性函数可以是未知的。在基于性能势的优化方程框架下,通过系统样本路径设计性能势函数的在线学习算法,进一步给出优化控制算法。在此基础上利用RBF神经网络的逼近特性,给出简化学习算法以减少计算量。最后给出的具体仿真结果和数据表明本文算法的有效性。

This paper concerns on optimization of non- linear stochastic system, which the non- linear function may be unknown. The on - line algorithm based on performance potential is designed from a sample path. The RBF network approach reduces the calculation. The results listed in this paper indicate the efficiency of the algorithm.