Abstract
A simple regulator having an unknown gain in the control path is considered. Information on the unknown gain can be obtained only by using the control as a probe. An optimal estimation procedure is developed using sufficient statistics. The optimal control was computed by Bellman's method of dynamic programming using a function space approximation. Numerical results for the value of stationary control are given. The results were computed on a machine having a 1000 word working store.

This publication has 4 references indexed in Scilit: