Worst case Dynamic Programming and its application to deterministic systems
- View All
In this paper, we investigate the numerical solution of discrete time, infinite horizon, stationary, discounted Dynamic Programming(DP) problem. We introduce the worst case variant of the standard DP operator and show that it retains the nice properties of the former.We apply this worst case formulation to deterministic systems in continuous space and show that for a class of such systems, any arbitrary degree of accuracy can be attained in the numerical solution to the DP problem. Finally we illustrate our results through a simple orbital dynamics example. © 2000 The American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
author list (cited authors)
Chakravorty, S., & Hyland, D. C.