This chapter presents a multiobjective approach for the vehicle routing problem with optional collections, whose objectives are the minimization of the route costs and of the not fulfilled collection- demands. It proposes a data structure that best suits the problem, neighborhood structures that exploit both goals of the problem and an algorithm that checks the feasibility of a solution with lower computational cost. To solve the problem, three metaheuristics are discussed: the multiobjective iterated local search (MOILS), NSGA-II and the e-Constrained method, which are applied to fourteen instances containing between 50 and 199 customers. The results indicate that the MOILS outperforms the other approaches, obtaining significantly better average values for coverage, hypervolume and cardinality over the set of used test problems.
|Translated title of the contribution||Multiobjective Vehicle Routing Problem with Optional Collections|
|Title of host publication||Meta-Heurísticas em Pesquisa Operacional.|
|Publication status||Published - 9 May 2013|