Optimal control with a cost of changing control

Abstract

In this paper, we consider a class of optimal control problems where the variation of the control variable is appended as a penalty term into the cost function to reduce the fluctuation of the control. A new computational approach is developed for solving this type of problems, based on control parametrization technique used in conjunction with the time scaling transform, the constraint transcription method and a smoothing technique. This computational method is supported by rigorous convergence analysis.