数学物理学报(英文版) ›› 1985, Vol. 5 ›› Issue (4): 399-413.
韩崇昭
Han Chongzhao
摘要: This paper gives a mathematical definition for the "caution" and "probing", and presents a decomposition theorem for nonlinear discrete-time stochastic systems. Under some assumptions, the problem of finding the closed-loop optimal control can be decomposed into three problems:the deterministic optimal feedback, cautious optimal and probing optimal control problems.