Gradient and Quasi-linearization Computational Techniques for Distributed Parameter Systems

Abstract

This work discusses the mathematical description of distributed parameter systems, the statement of a maximum principle for a particular class of systems and the problems associated with approximating distributed systems by discretization into lumped systems for computational purposes. The gradient and quusi-linearization methods have been applied, perhaps for the first time, to a class of nonlinear distributed parameter systems. Examples concerning the optimal control of non-linear distributed regulator problems are presented.