D~2G~2A: A Distributed Double Guided Genetic Algorithm for Max_CSPs

摘要

Inspired by the distributed guided genetic algorithm (DGGA), D~2G~2A is a new multi-agent approach, which addresses Maximal Constraint Satisfaction Problems (Max-CSP). GA efficiency provides good solution quality for Max_CSPs in one band and benefits from multi-agent principles reducing GA temporal complexity. In addition to that the approach will be enhanced by a new parameter called guidance operator. The latter allows not only diversification but also an escaping from local optima. D~2G~2A and DGGA are been applied to a number of randomly generated Max_CSPs. In order to show D~2G~2A advantages, experimental comparison is provided. As well, guidance operator is experimentally outlined in order to determine its best given value.