Adaptive multi-agent coordinating genetic algorithm

An adaptive multi-agent coordinating genetic algorithm is presented. This algorithm is composed of relatively independent agents under a coordinator, each agent being a computing unit based on simple genetic algorithms. The coordinator allocates the computing resources adaptively according to the output and growth-rate of each agent, so the limited computing resources are utilized much more reasonably. Agents collaborate through individual migrations and cross-area crossovers, individual migrations introduce new genes, cross-area crossovers produce better offspring individuals, thus the quality of individuals is improved. Testing result shows that this algorithm has very good convergence.