Improved Penalty Algorithm for Mixed Integer PDE Constrained Optimization (MIPDECO) Problems

07/15/2019

∙

by   Dominik Garmatter, et al.

∙

0

∙

share

Optimal control problems including partial differential equation (PDE) as well as integer constraints combine the combinatorial difficulties of integer programming and large-scale systems resulting from discretizing the PDE. A common solution strategy for such a problem would be Branch-and-Bound. In order to provide an alternative solution approach, especially in a large-scale context, this article investigates penalization techniques. Taking inspiration from a well-known family of existing Exact Penalty algorithms a novel Improved Penalty Algorithm is derived. Using a standard stationary model problem, the algorithm is then numerically investigated and compared with different solution strategies including a naive penalty approach and the Branch-and-Bound routine from CPLEX.