Introducing a linkage identification considering non-monotonicity to multi-objective evolutionary optimization with decomposition for real-valued functions

Kousuke Izumiya, Masaharu Munetomo

Abstract

In this paper, we propose MOEA/D-LIEM2 to optimize multiple real-valued objective functions which have complex interaction among variables by employing linkage identification. We compared the proposed algorithm with MOEA/D without linkage identification. As a result, we found that the proposed algorithm out-performs the original MOEA/D for problems consisting of two deceptive functions with linkage.
BibTeX
@inproceedings{izumiya2018introducing,
  author = {Kousuke Izumiya and Masaharu Munetomo},
  title = {{Introducing a linkage identification considering non-monotonicity to multi-objective evolutionary optimization with decomposition for real-valued functions}},
  booktitle = {GECCO (Companion)},
  year = {2018},
  doi = {10.1145/3205651.3205692},
}

Loading related content...

Figure