MULTIOBJECTIVE TABU SEARCH WITH MIXED INTEGERS AND REFERENCE POINT

Abstract

In this work we present a domain-independent Tabu Search approach for multiobjective optimization with mixed-integer variables. In this we investigate two aspects: domain-independence and applicability in optimization practice and focus our attention in problems that appear frequently in the real world, like logistic network (for example: multi-stage distribution networks problems, location-allocation problems, time-tabling problems); however, other classical problems were investigated, like: coverage set problem, partitioning set problem, multidimentional knapsack problem and shortest path problem. All these problems belong to the NP-hard class, with a great number of decision variables, containing a great number of heterogeneous constrains, presenting a challenge to find feasible solutions.

Keywords

• múltiples objetivos
• metaheurísticas
• búsqueda tabú