具有模型和实际差异的非线性系统最优控制算法及其收敛性(英文)

Convergence of an Algorithm for Optimal Control ofNonlinear ContinuousTime Systems with ModelReality Differences*

针对模型和实际之间的差异，提出了一种基于时变线性二次型问题的动态系统优化和参数估计集成的算法，该算法能逼近实际问题最优解．给出了该算法收敛的一个充分条件，分析了它的最优性．仿真例子说明了该算法的有效性和实用性．

This paper presents an extension algorithm of dynamic integrated system optimization and parameter estimation based on timevariant linearquadratic problem for nonlinear optimal control where modelreality differences exist.Convergence of the algorithm is inverstigated.A sufficient condition is derived for the convergence and optimality of the algorithm.It is shown that limit point of the algorithm solution satisfies the maximum principle for optimal control.The implementable conditions of the algorithm are emphasized,simulation example denotes the efficiency and the applicability of the algorithm.