Smoothing technique of nonsmooth Newton methods for control-state constrained optimal control problems

摘要

This paper is concerned with the numerical solution of optimal control problems subject to mixed control-state constraints with differential-algebraic equations. The necessary conditions arising from a local minimum principle are transformed into an equivalent nonsmooth equation in appropriate Banach spaces. We then develop a parametric smoothing Newton method for solving this nonsmooth equation. The global and local convergence is proposed under suitable settings. Numerical examples are presented to demonstrate the efficiency of our global smoothing technique.