Metaheuristics based on decision hierarchies for the traveling purchaser problem

In this paper we address the traveling purchaser problem, an NP-hard problem that generalizes the traveling salesman problem. We present several metaheuristics that combine genetic algorithms and local search. The genetic algorithms are induced by different hierarchic orderings of the decision making regarding the route and the acquisition of the items. Computational experiments were carried out with benchmark instances and the results show that the proposed metaheuristics are a suitable tool to solve high-dimensioned instances for which the exact methods do not provide solutions within a reasonable CPU time. For several instances, best new upper bounds for the optimum value of the objective function were obtained.