A computing technique for solving continuous and discrete time optimal control problems

Abstract

This paper presents the results of investigations on the application to optimal control problems of a new optimization technique based on Liapunov concepts. In this method, the actual search is carried out on a positive-definite Liapunov Type of Function (LTF) of the squared values of the gradients and other components of an augmented objective function. The distinguishing feature of this method is that the solution of all types of optimal control problems can be obtained by a single unconstrained optimization of the LTF. It is shown that the simple Gauss-Newton Least Square method is very well suited for this particular structure as the CPU time is lower than for the variable metric methods.