To slide or not to slide? Moving along fitness levels and preserving the gene subsets diversity in modern evolutionary computation
Michal W. Przewozniczek, Marcin M. Komarnicki · Proceedings of the Genetic and Evolutionary Computation Conference · 2023
Optimal Mixing (OM) is a mating operator employed by many state-of-the-art Genetic Algorithms (GAs). This paper identifies the sliding phenomenon defined as a serie of the so-called plateau moves. Sliding is a part of the original OM proposition. Although sliding seems to be a minor element of OM, we show that performing or avoiding it may significantly affect the effectiveness of OM-employing GAs. Therefore, we analyze the details of sliding pros and cons and propose the Autonomous Slide Deciding Algorithm (ASDA). ASDA analyzes the diversity of the population for a given subset of genes. Then, it decides, for a given mixing operation, sliding will be profitable or not. We show that using ASDA is greedily beneficial for two different state-of-the-art GAs. Additionally, we explain why ASDA deteriorates the effectiveness of the third considered OM-using GA.