Neighboring Optimal Method for the Solution of Optimal Control Problems

摘要

The solution to a general class of optimal control problems is derived using the neighboring optimal method. The problem is first transformed using the minimum principle of Pontryagin into a two-point boundary value problem. The solution to the boundary value problem is formulated as minimizing a terminal error function. Sensitivity coefficients are derived to provide a means, in conjunction with a linear search, of sequentually approaching the solution. The sensitivity coefficients are also used to derive a feedback control law. The neighboring optimal method can be used in fixed time problems, and problems with either free or fixed terminal state conditions. (Author)