摘要
许多现实世界中的优化问题都是多个目标的,而且是和时间因素有关的,抽象成数学模型就是动态的多目标优化问题,基于免疫遗忘概念和免疫应答的动态过程,提出了一种用于解决动态多目标优化问题的新的人工免疫系统算法-免疫遗忘动态多目标优化(IFDMO)算法.并采用了两集合覆盖这一评价参数,对算法进行了定量的描述.这一参数用于测量在每一个时间步骤得到的最优解向着Pareto-最优面的逼近程度.并将该算法与另外一种算法CSADMO进行了比较,CSADMO是最近提出的一种用于解决动态多目标优化问题的方法,CSADMO在保持所得前沿面的均匀性,多样性及向着Pareto-最优面的逼近性方面都体现出了很好的性质.实验结果表明,在每。时间步骤中,与CSADMO相比,IFDMO获得的解能更好的向着Pareto-最优面逼近,而且解得分布也更加均匀,范围也更加宽广.
Clonal Selection Algorithm for Dynamic Multiobjective Optimization (CSADMO) is a relatively new technique for finding or approximating the Pareto-optimal front every time there is a change in t for dynamic multiobjective optimization problems CSADMO has shown good performance in both the convergence and diversity of obtained solutions in comparison to another dynamic multiobjective optimization Algorithm: A Direction-Based Method (DBM). In this paper, based on the artificial immune system and the dynamic process of immune response, a new dynamic multiobjective optimization algorithm termed as Immune Forgetting dynamic multiobjective optimization (IFDMO) is proposed. Simulation results of the IFDMO on four test problems are compared with CSADMO and much better performance in both the convergence and diversity of obtained solutions of CSADMO is observed.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2006年第B07期205-209,共5页
Journal of Harbin Engineering University
关键词
人工
免疫系统
免疫遗忘
动态多目标优化
性能评价
artificial immune system
immune forgetting
dynamic multi-objective optimization
performance evaluation