Non-elitist Evolutionary Multi-objective Optimizers Revisited
This challenges a long-held belief in optimization research, showing non-elitist methods can be effective for specific high-dimensional problems.
This paper revisits non-elitist evolutionary multi-objective optimization algorithms (EMOAs) for bi-objective continuous optimization, finding that they perform significantly well on problems with many decision variables when using an unbounded external archive, contrary to conventional wisdom.
Since around 2000, it has been considered that elitist evolutionary multi-objective optimization algorithms (EMOAs) always outperform non-elitist EMOAs. This paper revisits the performance of non-elitist EMOAs for bi-objective continuous optimization when using an unbounded external archive. This paper examines the performance of EMOAs with two elitist and one non-elitist environmental selections. The performance of EMOAs is evaluated on the bi-objective BBOB problem suite provided by the COCO platform. In contrast to conventional wisdom, results show that non-elitist EMOAs with particular crossover methods perform significantly well on the bi-objective BBOB problems with many decision variables when using the unbounded external archive. This paper also analyzes the properties of the non-elitist selection.