A status report on the development of near optimal differential game feedback methods for realistic nonlinear problems

Two classes of techniques for generating feedback controls, based on zero-sum perfect information differential game theory, are reviewed. The first class makes use of highly simplified dynamics as an approximation to the actual problem dynamics in order to obtain feed-back controls in closed-form. The second class of techniques uses sophisticated computational algorithms to generate near-optimal feedback controls based on realistic high order nonlinear system dynamics. The advantages and shortcomings of both of these approaches are discussed, along with possible future developments in this area.