Linear Programming and Suboptimal Solutions of Dynamical Systems Control Problems

摘要

Linear programming is used to obtain a direct search procedure for parameter optimization in the numerical solution of dynamical systems control problems including those where state inequality constraints are presented. The procedure applies to optimal control problems which uses the Raileigh-Ritz method to obtain a suboptimal approximation, imposing a particular mathematical form to the control function, making it dependent upon a finite number of parameters. It is shown that state inequality constraints can be included without the need of using penalty function methods.