An iterative technique for the computation of the steady state gains for the discrete optimal regulator

摘要: … The purpose of this correspondence is to establish the validity of an algorithm for the discrete-time invariant linear regulator with infinite terminal time, which is the analog of the algorithm derived by Kleinman [5] for the continuous problem. Although the usual derivat.ion of the optimal control law is recursive in nature, the rate of convergence to the steady state is generally not very rapid, especially in the neighborhood of the steady state, as simple one-dimensional examples demonstrate. In contrast, the algorithm presented here converges in a …