New concave penalty functions for improving the Feasibility Pump

Abstract

Mixed-Integer optimization represents a powerful tool for modeling manyoptimization problems arising from real-world applications. The Feasibilitypump is a heuristic for finding feasible solutions to mixed integer linear problems. In this work, we propose a new feasibilitypump approach using concave nondifferentiable penaltyfunctions for measuring solution integrality. We present computational results on binaryMILP problems from the MIPLIB librarysho wing the effectiveness of our approach.