基于NSGA-II的改进多目标遗传算法  

Improved multi-objective genetic algorithm based on NSGA-II

作　　者：陈小庆[1] 侯中喜[1] 郭良民[1] 罗文彩[1] 

机构地区：[1]国防科学技术大学航天与材料工程学院,湖南长沙410073

出　　处：《计算机应用》2006年第10期2453-2456,共4页journal of Computer Applications

基　　金：国家863计划项目(2005AA756050)

摘　　要：在已有多目标优化算法(NSGA-II)研究和分析的基础上,为加快收敛速度,提高收敛精度,设计了新的初始筛选机制,改进了交叉算子的系数生成,提出了更为合理的排挤机制。通过典型应用函数的计算测试,结果表明:上述改进不仅具有较高的计算效率,而且能够得到分布更为合理的解,且能保持解的多样性分布。Based on the study and analysis of NSGA-Ⅱ algorithm, a new initial screening mechanism was designed, coefficient generating of crossover arithmetic operator was improved and more reasonable crowding mechanism was proposed. In this way, convergence was speeded up and its precision was improved. The testing results by representative applied functions show that with the improvements higher computational efficiency and more reasonable distributed solution can be obtained, and diversified distribution of the solutions can be maintained.

关 键 词：多目标优化 遗传算法 排挤机制 交叉算子 初始种群 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP301.6[自动化与计算机技术—控制科学与工程]