THE POWER OF NATURAL EVOLUTION IN SOLVING DISTRIBUTED CONSTRAINT OPTIMIZATION PROBLEMS

As a popular search method, genetic algorithm (GA) has been applied to a wide range of optimization problems. This paper proposes GA-DCOP, a novel incomplete distributed algorithm which exploits GA's powerful global search ability to solve distributed constraint optimization problems. Concretely, within the DCOP framework, each agent is given the authority to conduct genetic operations independently. The proposed algorithm is evaluated experimentally and compared to other DCOP algorithms. The experimental results demonstrate that GA-DCOP outperforms several state-of-the-art algorithms in either solution quality or in simulated time.