Solving constraint satisfaction problems by using coevolutionary genetic algorithms

In this paper, Coevolutionary Genetic Algorithm for solving Constraint Satisfaction Problems (CSPs) is proposed. It consists of two Genetic Algorithms (GAs): a traditional GA and another GA to search for good schemata in the former GA. These GAs evolve in two levels, i.e., phenotype-level and schema-level, and affect with each other through genetic operations. To search for solutions effectively, we devise new genetic operator by utilizing search mechanism of solution synthesis approach used in CSP community. Computational results on general CSPs confirm the effectiveness of our approach.