Numerical methods for nonlinear optimal control problems: Application to abnormal problems

AbstractNew, efficient, accurate numerical methods are given for general classes of optimal control problems. It is seen that these results satisfy an a priori maximum pointwise component error of O(h2) with a Richardson error of O(h4). The foundation for these results is the corresponding results for the m-dependent-variable problem in the calculus of variations and the use of multipliers to convert the optimal control problems to the calculus of variations setting