Maintaining genetic diversity in fine-grained parallel genetic algorithms by combining cellular automata, Cambrian explosions and massive extinctions

This paper describes an evolutionary algorithm (EA) which combines cellular automata, Cambrian explosions and massive extinctions ideas in order to maintain diversity and automatically determine the population's size of the EA. Individuals are organized in a two-dimensional grid (2-dimensional cellular automaton surface) and are considered active or inactive according to the cellular automaton state. The individual state is updated according to the cellular automaton state rules at each step (iteration) of the evolutionary process. Only active individuals are subject to evolution by applying one of the genetic operators and considering just their active neighbors (when multiple parents are required). Depending on the total number of active individuals, a Cambrian explosion or a massive extinction operation is applied, in a random fashion to control the size of the population. We presented a novel genetic diversity analysis using a hierarchical clustering to examine individuals genotype and identify natural population taxonomies. Our experiments show that the proposed scheme is able to maintain diversity and find near optimal solutions in an appropriated number of fitness evaluations.