An Evolutionary Multi-Agent Algorithm Applied to the Multi-Objective Optimization Problem
MI Zhong-chun · Jisuanji fangzhen · 2009
An evolutionary multi-agent algorithm is proposed to search Pareto solutions for Multi-objective optimization problem.Every single objective is assigned to a single agent.Every agent has its colony and takes the assigned objective as the estimation function for the evolution.The estimation of distribution for elite solutions of every agent's colony will be obtained.The distributions will be synthesized through communication and based on which the next generations will be produced.Tests show that solutions can converge to Pareto set within fewer generations.Pareto set can be obtained through using this evolutionary algorithm efficiently.