An Optimal Control Algorithm Based on an Initial Costate Search

摘要

An algorithm is presented for solving a wide class of optimal control problems. The algorithm is based on the reformulation of the associated two-point boundary value problem as a multi-dimensional minimization problem. This secondary problem is solved using the method of Davidon-Fletcher-Powell. Two techniques are considered for solving the one-dimensional minimization problem implicit in the Davidon-Fletcher-Powell method; the first is a polynomial interpolation technique, and the second a scheme based on the Fibonacci search. A variety of example problems are solved to demonstrate both the scope and the effectiveness of the algorithm. (Author)