Optimal control for discrete stochastic system with conditional Markov structure
具有条件马尔科夫结构的离散随机系统最优控制

Based on Bellman stochastic nonlinear dynamic programming, we develop an optimal control algorithm for the discrete stochastic system with conditional Markov structure. By making use of characters of stochastic variable structure system, we simplify the optimal control algorithm, and approximate the posterior probability density function by a Gaussian function. The optimal control algorithm for a class of linear discrete stochastic systems with conditional Markov structure is derived finally.