Controlled Local Search for the Hybridization of Evolutionary Algorithms in Multi-Objective Optimization
1 : Laboratoire de Technologies de l'Information (LTI)
Ecole Nationale des Sciences Appliquées- El Jadida (ENSA-J) Université Chouaib Doukkali 24002 El Jadida, Morocco -
Morocco
The hybridization of MOO, which is proposed in this paper, has three specific characteristics: (i) the evolutionary search is based on the Pareto MOEA (i.e. NSGA, NSGA-II, SPEA, SPEA-2 and SPEA2+); (ii) local search does not use objective scaling, instead it uses a Multi-Objective Local Search in two different forms (diversity-MOLS to have large-sized Pareto front and convergence-MOLS to come closer to optimal Pareto front); and (iii) different hybridization techniques are discussed/implemented and evaluated.
