A Multi-Agent Self-Adaptive Multi-Objective Genetic Algorithm

The agent technology and genetic algorithms are integrated and is applied to solve multi-objective optimization problem. An agent in this algorithm represents a candidate solution to the multi-objective optimization problem. Agent lives in the grid environment and it possesses own local space called the neighborhood. In the neighborhood, an agent can compete and collaborate with other agents to achieve the purpose of gene exchange and evolution. Agent also possesses some knowledge of the environment and can learn itself while evolving, in order to adapt itself to the environment better and enhance its viability. A new multi-objective genetic algorithm based on Multi-Agent Self-Adaptive Genetic Algorithm(MASAGA) is proposed, in which the evolution parameters are adjusted adaptively in the evolutionary process and a new selection operator is used to select individual. By adjusting the crossover and mutation parameters in the evolutionary process it can improve the accuracy and convergence speed of the algorithm. Several benchmark functions are used to test the performance of the algorithm and the simulation results indicate that the multi-objective evolutionary algorithm based on MASAGA has a better performance. The algorithm can converge to the Pareto solutions quickly, and has a good diversity compared with NSGA-II.