A unified approach to computational methods of nonlinear optimal control problems with possible jumps in states

A unified approach for numerical solutions of nonlinear optimal control problems is proposed. First, a canonical form of optimal control problems is defined, which includes a variety of control problems, such as time optimal ones, design parameter ones, initial state constrained ones, state discontinuous ones, and so on. Then, based on the algorithms of the Bolza type, the algorithm for numerical solutions of the canonical problems is derived. Lastly, the usefulness of the canonical setting of optimal control problems is demonstrated through some numerical simulations. It should be noted that the canonical problem is considered to be one of the Bolza type problems and so powerful algorithms developed for those problems are available.