机构地区: 青岛科技大学计算机与化工研究所
出 处: 《计算机工程与应用》 2004年第32期85-88,共4页
摘 要: 在传统的微粒群优化算法的基础上,提出了一种基于动态Pareto解集的求解多目标规划问题的方法。Pareto解集在每次迭代过程中进行动态更新和信息共享,在加入新产生的Pareto近似最优解同时去除解集中已经不是Pareto解的数据,每个个体随机地与Pareto解集中的结果进行信息交换,从而保证在快速找到Pareto解的同时保持多样性。并通过三个标准的测试函数证明了算法的有效性。 A new dynamic Pareto warehouse-based Particle Swarm Optimization(DPW-PSO)algorithm is developed based on the traditional PSO algorithm.Data in the Pareto warehouse is dynamically updated and the information is shared by all the individuals in every iteration cycle.New Pareto results are added to the warehouse and the pseudo Pareto results are deleted at the same time ,individuals can exchange randomly the information with all the data in the Pareto warehouse to guarantee finding the Pareto results quickly and variously.Three standard test functions are illustrated to show the validity of the method.
领 域: [自动化与计算机技术] [自动化与计算机技术]