A Sufficient Condition for the Temporal Decomposition of Optimum-Control Problems for Linear Stochastic Systems with Restricted Controllers.
Theory and derivations for transforming an optimum-control problem for linear stochastic systems with restricted controllers into an ordinary estimation problem. A sufficient condition is obtained to define a subset of estimation problems, and hence control problems, for which suboptimizing at each time epoch is an overall optimum strategy. The results of this analysis are applied to example control problems and relationships to previous findings are indicated. 22 pp. Refs. (KB)