Nonlinear Optimal Control with Tensors: Some Computational Issues

Researches in the last six years have made possible a new computational approach to problems of nonlinear optimal feedback control. Recursive equations for the terms in expansions of optimal value functions and optimal feedback control laws have been determined, and have been parameterized in terms of the algebraic tensor. The first numerical results based upon these new theories were reported by O´Sullivan and Sain in 1984. This paper discusses some computational issues which arise in such calculations.