Technical Note—A Note on “Myopic Solutions of Markovian Decision Processes and Stochastic Games”
Operations Research · 1985 · Volume 33 · Issue 6
Abstract
The aim of this note is to correct some oversights in a 1981 paper by M. J. Sobel on myopic policies. We show by example that somewhat stronger assumptions than Sobel's must be imposed on a Markovian decision model to ensure the optimality of myopic policies. We offer alternatives for augmenting or altenng Sobel's assumptions so that they will be sufficient for myopic policies to be optimal from every initial state.
Keywords
116 Markov decision processes, myopic optimal policies
Enable JavaScript for citation tools and additional paper interactions.