摘要
在Evo-Ant算法的基础上提出了多目标的算法,即利用Evo-Ant算法来产生新的解,并利用一个额外的存储空间来存放Pareto候选解,用新产生的解来更新Pareto候选解,消除被支配的解,依次循环,从而得到近似的Pareto解.为了验证演化蚁群算法,采用2种测试手段:一种是Solomon的测试数据;另一种是在仿真环境下的测试.实验结果表明该算法很具有竞争能力.
A new MOPs algorithm, named MEvo-Ant（multi-objective evolutionary ant algorithm）, is proposed for the DVRPs（dynamic vehicle routing problem ）. In the MEvo-Ant algorithm an archive is used to storage the candidate Pareto solutions while the Evo-Ant algorithm is employed to generate new solutions and update the archive to eliminate the dominated solutions. With the iterations the solutions in the candidate Pareto set are approach to the true Pareto solutions. In this paper two methods are used to evaluate the MEvo-Ant, the one is used the Solomon testing data, and the other is used a simulator. The experiment results show that the MEvo-Ant algorithm is effective to the DVRPs.
出处
《武汉大学学报（理学版）》
CAS
CSCD
北大核心
2007年第5期571-575,共5页
Journal of Wuhan University:Natural Science Edition
基金
国家自然科学基金资助项目(60301009)
武汉市青年科技晨光计划项目(20065004116-03)
关键词
演化蚁群算法
动态车辆路径问题
多目标优化
evolutionary ant algorithm
dynamic vechicle routing problem （DVRP）
multi-objective optimization
