On the construction of ℝà-optimal strategies in partially observed MDPs

The purpose of the paper is to give a survey of methods, partly derived by the author in joint work with other researchers, concerning the problem of constructing •-optimal strategies for partially observable MDPs. The methods basically consist in transforming the problem into one of approximation: Starting from the original problem a sequence of approximating problems is constructed such that: (i) For each approximating problem an optimal strategy can actually be computed. (ii) Given •>0, there exists an approximating problem such that the optimal strategy for the latter is •-optimal for the original problem.