The success of local search techniques in the solution of combinatorial op-timization problems has motivated their incorporation into multi-objectiveevolutionary algorithms, giving rise to the so-called multi-objective memeticalgorithms (MOMAs). The main advantage for adopting this sort of hy-bridization is to speedup convergence to the Pareto front. However, the useof MOMAs introduces new issues, such as how to select the solutions to whichthe local search will be applied and for how long to run the local search engine(the use of such a local search engine has an extra computational cost).Our work involves to design a new MOMA, in order to decrease the runningtime of its computation, but without lose the offered advantages. In thisthesis, we propose an IGD+-based local search engine (IGD+ is consideredas reference-based performance indicator), which adopts a novel clusteringtechnique for splitting the objective space into sub-regions.
