• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2011, Vol. 33 ›› Issue (3): 141-145.doi: 10.3969/j.issn.1007130X.2011.

• 论文 • 上一篇    下一篇

PBIL算法在组合优化问题中的应用研究

袁利永1,倪应华2,金炳尧3,马永进1   

  1. (1.浙江师范大学数理与信息工程学院,浙江 金华 321004;2.浙江师范大学行知学院,浙江 金华 321004;
    3.浙江师范大学信息传播实验教学中心,浙江 金华 321004)
  • 收稿日期:2010-03-11 修回日期:2010-05-29 出版日期:2011-03-25 发布日期:2011-03-25
  • 作者简介:袁利永(1978),男,浙江嵊州人,讲师,研究方向为进化计算和机器学习。倪应华(1977),男,浙江海盐人,讲师,研究方向为计算机辅助技术和多媒体技术应用。金炳尧(1964),男,浙江兰溪人,硕士,教授,研究方向为模式识别和进化计算。
  • 基金资助:

    2008年度浙江省教育厅项目(Y200805671)

Application of the PBIL Algorithm to the Combinatorial Problem

YUAN Liyong1,NI Yinghua2,JIN Bingyao3,MA Yongjin1   

  1. (1.School of Mathematics Physics and Information Engineering,Zhejiang Normal University,Jinhua 321004;
    2.Xinzhi School,Zhejiang Normal University,Jinhua 321004;
    3.Information Dissemination Experimental Teaching Center,Zhejiang Normal University,Jinhua 321004,China)
  • Received:2010-03-11 Revised:2010-05-29 Online:2011-03-25 Published:2011-03-25

摘要:

基于群体的增量学习(PBIL)算法有效结合了遗传算法和竞争学习的优点,运行过程简单,解决问题快速准确。本文提出将PBIL算法应用于求解CMN组合优化问题,以物流中心选址优化问题为例,介绍了基于PBIL求解CMN组合优化问题的一般方法,提出了针对此类问题的个体产生算法。为了提高算法的收敛速度和寻优能力,提出了基于当代最优解与历代最优解比较结果的概率学习加速方法。最后,通过实验仿真验证了上述改进的有效性。

关键词: PBIL, 组合优化, 进化计算, 基因算法, 物流中心选址

Abstract:

PBIL combines the features of genetic algorithms(GA) and competitive learning in an efficient way, which has the advantage of simple execution process, quick and accurate solutions to problems. In this paper, the PBIL algorithm is applied to solving combinatorial optimization problems. Using the logistics center location as an example,we  illustrate a general method of solving the combinatorial optimization problems based on PBIL.A new algorithm for producing individuals for such problems is proposed. In order to improve the convergence speed and search capability, an acceleration method of probability learning is put forward based on the comparison of contemporary optimal solution and the successive optimal solution. Finally, the effectiveness of improvement is verified through simulation experiments.

Key words: PBIL;combinatorial optimization;evolutionary computation;genetic algorithm;logistics center location