基于统计线性化的随机非线性微分对策逼近最优策略  

Near Optimal Strategy for Nonlinear Stochastic Differential Games Based on the Technique of Statistical Linearization

作　　者：张平[1,2] 方洋旺[2] 惠晓滨[2] 刘新爱[1] 李亮[1] 

机构地区：[1]中国人民解放军第95889部队 [2]空军工程大学航空航天工程学院

出　　处：《自动化学报》2013年第4期390-399,共10页Acta Automatica Sinica

基　　金：国家自然科学基金(60874040)资助~~

摘　　要：针对二人零和随机非线性微分对策问题,利用统计线性化技术并提出一种新的逼近策略控制方法.通过求解具有统计线性化参数的Riccati微分方程得到逼近控制策略,该Riccati微分方程与一般线性系统的Riccati微分方程具有明显的区别;同时对控制量受约束情形的控制策略也进行了求解;最后通过仿真验证了所得结论的正确性.A novel solution for a class of nonlinear zero-sum stochastic differential games is given based on the technique of statistical linearization. The near optimal feedback strategies are derived by solving the statistical state dependent Riccati equation, which is significantly different from the Riccati equation of linear systems. The case of strategy with bound limitation is also investigated. An example is given to illustrate the application of the theory.

关 键 词：微分对策 统计线性化 随机非线性系统 逼近最优策略 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]