参数未知系统的多模型对偶控制算法

Dual Control Algorithm for System with Parameters Unknown

论文针对参数未知的随机系统提出一种多模型对偶控制算法。首先使用多个子模型代替参数未知的系统模型,然后以传统的线性二次型高斯控制(LQG)为理论依据,引入后验概率,在多个模型中筛选出综合的参数模型,最后利用该模型对系统进行控制,获得一种新的多模型对偶控制策略。仿真结果表明了该文算法的有效性。

A multi-model dual control algorithm is presented for stochastic system with parameters unknown. Multiple models are used to describe the system model first, and then the integrate model is identified from multiple models via joining the posterior probability based on the LQG control theory. Finally, the integrate model is uesd to control the system. Simulation results illustrate the effectiveness of the algorithm.