Partially Observable Control Problems with Compulsory Shifts of the State

Runggaldier, W.J. & Stettner, L. (1992). Partially Observable Control Problems with Compulsory Shifts of the State. IIASA Working Paper. IIASA, Laxenburg, Austria: WP-92-034

Preview

Text WP-92-034.pdf Download (577kB) | Preview

Abstract

Stochastic control problems with partial state observation and the long-run average cost criterion are among the most difficult dynamic stochastic optimization problems and almost nothing has so far appeared in the literature concerning their solution. On the other hand many problems in Engineering, Operations Research, and the Economic and Social Sciences can be modelled as problems of the above type. In the present paper we study conditions under which the filtering process associated with the partially observed state process has a unique invariant measure and describe ways to approximate it. We finally discuss the applications of these results to the construction of nearly optimal controls.

Actions (login required)

View Item