Preserving Diversity through Diploidy and Meiosis for Improved Genetic Algorithm Performance in Dynamic Environments

Abstract

Genetic algorithms have been applied to a diverse field of problems with promising results. Using genetic algorithms modified to various degrees for tackling dynamic problems has attracted much attention in recent years. The main reason classical genetic algorithms do not perform well in such problems is that they converge and lose their genetic diversity. However, to be able to adapt to a change in the environment, diversity must be maintained in the gene pool of the population. One approach to the problem involves a diploid representation of individuals. Using this representation with a dynamic dominance map mechanism and meiotic cell division helps preserve diversity. In this paper, the effects of using diploidy and meiosis with such a dominance mechanism are explored. Experiments are carried out using a variation of the 0-1 knapsack problem as a testbed to determine the effects of the different aspects of the approach on population diversity and performance. The results obtained show promising enhancements.