A system for the numerical solution of constrained optimal control problems

In this paper an optimal control problem with Ordinary Differential Equations (ODE’s) and state and control constraints is considered. The present approach is to convert the control problem into a finite-dimensional nonlinear program by parametrization of the control and discretization of the ODE. The controls will be approximated by piecewise constant or piecewise linear functions and characterized by a finite set of parameters which includes function values and switching points. An algorithm with an efficient procedure for computing gradients and its implementation is presented. The nonlinear program is then solved by sequential quadratic algorithms. Some numerical results are presented.