摘要
针对约束多目标优化问题,提出了一种新型的约束多目标优化算法。该算法采用了一种新型约束处理方式,先通过约束违反门限截取种群再依据约束与目标函数值针对不同情况实现对个体的优劣划分。本算法将差分进化与免疫克隆机制相融合,既利用了差分进化从全局角度进行搜索的特点,又利用了免疫克隆机制从优秀个体出发进行局部再寻优搜索的优点,扩大了算法搜索的广度与深度。测试结果表明该算法相比快速非支配排序遗传算法(NSGA-II)具有非常优秀的收敛性与分布性。将提出的算法应用于实际的汽油调合优化中,进一步验证了算法的有效性,可有效减少成本,提高产品质量。
This paper proposes a novel algorithm for solving constrained multi-objective optimization problems. The proposed approach adopts a new kind of constraint handling method which firstly cut the population with the threshold of constraint violation and then divide the individuals according to the different situations, constraint violation and the objective function values. In addition, a hybrid of differential evolution and immune clonal mechanism is introduced in this paper, in order to search in global by differential evolution and search in local based on the elite by immune clone. The experimental results show that this algorithm compared with NSGA-II has good convergence and distribution. Finally, the proposed algorithm is used in the optimization of actual gasoline blending operation, which verify the efficiency of the proposed method, and provide a new way to reduce cost and improve the quality of products for the company.
出处
《华东理工大学学报(自然科学版)》
CAS
CSCD
北大核心
2013年第3期311-318,共8页
Journal of East China University of Science and Technology
基金
国家自然科学基金项目重点基金(U1162202)
国家"863"计划项目(2012AA040307)
国家自然科学基金项目(61174118)
中央高校基本科研费专项资金
上海市重点学科建设项目(B504)
关键词
约束处理
差分进化
免疫克隆
多目标
汽油调合
constraint handling
differential evolution
immune clone
multi-objective
gasoline blend