Computational aspects of infeasibility analysis in mixed integer programming

Abstract The analysis of infeasible subproblems plays an important role in solving mixed integer programs (MIPs) and is implemented most major MIP solvers. There are two fundamentally different concepts to generate valid global constraints from subproblems: conflict graph dual proof analysis. While detects sets contradicting variable bounds implication graph, derives linear the LP’s unboundedness. main contribution this paper twofold. Firstly, we present three enhancements analysis: presolving via cancellation, strengthening by applying rounding functions, a filtering mechanism. Further, provide comprehensive computational study evaluating impact every presented component regarding Secondly, presents first combined approach that uses both simultaneously within single solution process. All experiments carried out on general instances standard public test set Miplib 2017; algorithms have been non-commercial solver commercial .