不确定随机系统两步最小方差对偶控制  

Two-Step Minimization Variance Dual Control of Uncertainty Stochastic System

作　　者：高振斌[1] 

机构地区：[1]西安财经学院统计学院,陕西西安710100

出　　处：《哈尔滨理工大学学报》2014年第5期35-39,共5页Journal of Harbin University of Science and Technology

基　　金：陕西省教育厅专项科研计划项目(12JK0548);全国统计科研项目(LX2011LY052)

摘　　要：针对参数不确定随机系统,为寻求满足对偶性质的控制策略,将求解最优控制律的多步最小方差性能指标,用两步最小方差性能指标代替,不确定参数在有限模型集中取值;根据确定性等价原理,求取各模型的两步最小方差控制律,经过后验概率加权,得到系统的两步次优自适应对偶控制律.运用Matlab软件进行仿真试验,结果表明:该算法能有效地解决不确定随机系统估计与控制的矛盾,且算法简单、可行.For the stochastic system with unknown parameter,so as to achieve the control strategy that processing dual property,the multistep minimization variance performance index which can be used to solve the optimal control law is replaced with the two step minimization variance performance index.The uncertain parameter exist in the finite model set.In term of certainty equivalence principle,solving the two step minimization variance control law for every model,using posterior probability to weight,the two steps suboptimal adaptive dual control law is obtained.Matlab software is used to the simulating test,the results show that the algorithm can efficiently tackle the contradiction between the estimation and control and the method is simple and feasible.

关 键 词：不确定 随机系统 确定性等价原理 对偶控制 后验概率 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]