Improved random one-bit climbers with adaptive ε-ranking and tabu moves for many-objective optimization

Abstract:

Multi-objective random one-bit climbers (moRBCs) are one class of stochastic local search-based algorithms that maintain a reference population of solutions to guide their search. They have been shown to perform well in solving multi-objective optimization problems. In this work, we analyze the performance of moRBCs when modified by introducing tabu moves. We also study their behavior when the selection to update the reference population and archive is replaced with a procedure that provides an alternative mechanism for preserving the diversity among the solutions. We use several MNK-landscape models as test instances and apply statistical testings to analyze the results. Our study shows that the two modifications complement each other in significantly improving moRBCs' performance especially in many-objective problems. Moreover, they can play specific roles in enhancing the convergence and spread of moRBCs. © 2011 Springer-Verlag.

Año de publicación:

2011

Keywords:

Fuente:

scopus