On Infinite Stage Decision Processes

This paper extends the author’s [8] work on decision process, developed in the notation of Denardo [5], from the finite state (nonstationary) case to the infinite stage (nonstationary) case. The paper focuses on “structured” policies and strategies, and provides conditions under which such a policy (strategy) is e-optimal and/or optimal. Existence and “subfiniteness” issues are resolved in the face of unbounded returns.