一种用于多目标组合优化的三阶段混合蛙跳框架
作者:
作者单位:

1. 南京信息工程大学 自动化学院;2. 南京信息工程大学 江苏省大气环境与装备技术协同创新中心;3. 南京信息工程大学 江苏省大数据分析技术重点实验室

作者简介:

通讯作者:

中图分类号:

TP301.6

基金项目:

国家自然科学基金(61502239);江苏省自然科学基金(BK20150924),‘江苏省青蓝工程’资助。


A Three Phases Shuffled Frog Leaping Framework for Multi Objective Combinatorial Optimization
Author:
Affiliation:

1. School of Automation,Nanjing University of Information Science and Technology;2. Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology;3. Jiangsu Key Laboratory of Big Data Analysis Technology

Fund Project:

National Natural Science Foundation of China(61502239);Natural Science Foundation of Jiangsu Province(BK20150924),supported by "Qing Lan Project of Jiangsu Province of China".

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    提出一种用于求解多目标组合优化问题的三阶段混合蛙跳框架。该框架采用阶段化、模块化的设计思想,将种群的进化过程分为快速收敛、探索扩展、极值挖掘三个阶段。在快速收敛阶段,迅速定位Pareto前沿,使整个群体快速地聚集在前沿附近;在探索扩展阶段,进一步提升解的精度并让种群均匀地分布在前沿上;在极值挖掘阶段,搜寻各目标上的边界极值,增强分布性能。对于不同阶段的不同模块,采用不同的策略以提升框架的求解性能。所提框架对于多目标组合优化问题具有良好的通用性,在解决不同类型的问题时仅需设计相应的编码方式、个体生成算子和约束处理机制。采用经典的多目标背包问题作为测试问题,与五种已有算法进行对比,结果表明,所提框架具有良好性能,基于该框架设计的混合蛙跳算法具有更好的收敛性与分布性。

    Abstract:

    A three phases shuffled frog leaping framework for multi objective combinatorial optimization is proposed. The framework adopts the idea of phasing and modularization, and divides the evolutionary process of population into three phases: rapid convergence, exploration and expansion, exploit extremum. In the rapid convergence phase, the Pareto front is quickly positioned so that the whole population quickly gathers near the front. In the exploration and expansion phase, the accuracy of the solution is further improved and the population is evenly distributed on the frontier. In the exploit extremum phase, the boundary extremum on each objective is searched to enhance the distribution performance. For different modules in different stages, different strategies are adopted to improve the solution performance of the framework. The proposed framework has good generality for multi-objective combinatorial optimization problems. When solving different types of problems, only the corresponding coding mode, individual generating operator and constraint processing mechanism need to be designed. The classical multi-objective knapsack problem is used as the test problem, and compared with five existing algorithms, the results show that the proposed framework has good performance, and the hybrid frog hop algorithm based on the framework has better convergence and distribution.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-11-22
  • 最后修改日期:2021-02-03
  • 录用日期:2021-02-10
  • 在线发布日期: 2021-03-03
  • 出版日期: