抄録
For state inequality constrained optimal control problems, we have discontinuities in the adjoint variables at the junctions of unconstrained and constrained arcs. Thus it is very difficult to obtain the optimal solutions to these problems.
The purpose of this paper is to show that SUMUP (Sequential Unconstrained Method Using Penalty) for nonlinear programming can be extended to optimal control problems with state constraints.
The penalty function used in SUMUP is the generalized form of the functions for interior and exterior penalty methods and contains functions corresponding to Lagrange multipliers.
The computational features in this method are as follows:
1. We need not begin with a trajectory interior to the constraint set.
2. The integration of the adjoint equation does not become inaccurate.
3. This method can approach the optimal solution from the interior or exterior of the constraint set by choosing the functions corresponding to Lagrange multipliers.
An example is solved and the result presented.
