Statistical Comparisons of Classifiers over Multiple Data Sets
The Journal of Machine Learning Research
Multiobjective optimization problems with complicated Pareto sets, MOEA/D and NSGA-II
IEEE Transactions on Evolutionary Computation
A study of the parallelization of the multi-objective metaheuristic MOEA/D
LION'10 Proceedings of the 4th international conference on Learning and intelligent optimization
Multiobjective evolutionary algorithms: a comparative case studyand the strength Pareto approach
IEEE Transactions on Evolutionary Computation
MOEA/D: A Multiobjective Evolutionary Algorithm Based on Decomposition
IEEE Transactions on Evolutionary Computation
EuroGP'13 Proceedings of the 16th European conference on Genetic Programming
Hi-index | 0.00 |
MOEA/D is a multi-objective optimization algorithm based on decomposition, which consists in dividing a multi-objective problem into a number of single-objective sub-problems. This work presents two variants, called pMOEA/Dv1 and pMOEA/Dv2, of a new parallel model of MOEA/D that have been developed under the observation that different sub-problems may require different computational effort, and thus, demand different number of evaluations. Our interest in this paper is to analyze whether the proposed models are able of outperforming the MOEA/D in terms of the quality of the computed fronts. To cope with this issue, our proposals have been evaluated using a benchmark composed of eight problems and the obtained results have been compared against MOEA/D-DE, an extension of the original MOEA/D where new individuals are generated by an operator taken from differential evolution. Our experiments show that some configurations of pMOEA/Dv1 and pMOEA/Dv2 have been able to compute fronts of higher quality than MOEA/D-DE in many of the evaluated problems, giving room for further research in this line.