Improved Penalty Algorithm for Mixed Integer PDE Constrained Optimization (MIPDECO) Problems
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.
READ FULL TEXT