P. Amaral, P. Barahona
Infeasibility analysing can be quite difficult and complex even in the case of linear systems. Understanding conflicting constraints is helpful but is in general not enough for closing the subject. In general the next step is to take some action leading to a feasible system, and the only two options are removing or transforming the constraints. Both practices require additional information and have an unneglectable impact on the optimal value. In this presentation we propose a tool for infeasible LP, to guide the decision maker towards the most adequate strategy for dealing with the infeasibility. We present a mathematical formulation for the ranking of the optimal values and solutions among all feasible subsets of constraints, that is, to find (feasible) subsets of constraints that yield the K-best optimal values (K-Best Feasible Clusters). This practical and easily interpretable information can be crucial for deciding which constraints to drop from the original infeasible model.
Palabras clave / Keywords: infeasibility analysis, feasible clusters, repair of infeasible LP
Programado
Sesión bilateral SEIO-APDIO: Continuous Optimization (Organizadores: Tatiana Tchemisova y Margarita Rodríguez Álvarez)
1 de junio de 2018 12:30
Sala 6
