Adaptive-Q Application to Nonlinear Systems

Abstract

For nonlinear plants there has evolved both open-loop and closed-loop optimal control theory. Optimal feedback control for very general plants and indices is very new and appealing, see Elliott et al. (1994), but requires infinite dimensional controllers designed using off-line infinite dimensional calculations. Optimal open-loop nonlinear control methods are considered very elegant in theory, but lack robustness in practice, see Sage and White (1977). Can adaptive- Q methods somehow bridge the gap between these elegant theories and result in practical feedback controllers that have most of the benefits of both approaches without the disadvantages of each? We proceed as in the work of Imae, Irlicht, Obinata and Moore (1992) and Irlicht and Moore (1991).