Robust Optimization by ε -Ranking on High Dimensional Objective Spaces

Authors:
Hernán Aguirre;Kiyoshi Tanaka
Affiliations:
Fiber-Nanotech Young Researcher Empowerment Program, and Faculty of Engineering, Shinshu University, Nagano, Japan 380-8553;Faculty of Engineering, Shinshu University, Nagano, Japan 380-8553
Venue:
SEAL '08 Proceedings of the 7th International Conference on Simulated Evolution and Learning
Year:
2008

Citing 5
Cited 1

Multi-Objective Optimization Using Evolutionary Algorithms

Multi-Objective Optimization Using Evolutionary Algorithms
Evolutionary Algorithms for Solving Multi-Objective Problems

Evolutionary Algorithms for Solving Multi-Objective Problems
Combining convergence and diversity in evolutionary multiobjective optimization

Evolutionary Computation
Conflict, harmony, and independence: relationships in evolutionary multi-criterion optimisation

EMO'03 Proceedings of the 2nd international conference on Evolutionary multi-criterion optimization
Selection, drift, recombination, and mutation in multiobjective evolutionary algorithms on scalable MNK-Landscapes

EMO'05 Proceedings of the Third international conference on Evolutionary Multi-Criterion Optimization

Many-Objective Optimization by Space Partitioning and Adaptive ε-Ranking on MNK-Landscapes

EMO '09 Proceedings of the 5th International Conference on Evolutionary Multi-Criterion Optimization

Quantified Score

Hi-index	0.00

Visualization

Abstract

This work proposes a method to fine grain the ranking of solutions after they have been ranked by Pareto dominance, aiming to improve the performance of evolutionary algorithms on many objectives optimization problems. The re-ranking method uses a randomized sampling procedure to choose, from sets of equally ranked solutions, those solutions that will be given selective advantage. The sampling procedure favors a good distribution of the sampled solutions based on dominance regions wider than conventional Pareto dominance. We enhance NSGA-II with the proposed method and test its performance on MNK-Landscapes with up to M = 10 objectives. Experimental results show that convergence and diversity of the solutions found can improve remarkably on 3 ≤ M ≤ 10 objectives problems.