Journal of Systems Engineering and Electronics ›› 2009, Vol. 31 ›› Issue (7): 1642-1646.

• 系统工程 • 上一篇    下一篇

小生境免疫算法解决作业车间调度问题

赵良辉1, 杨海东2   

  1. 1. 五邑大学管理学院, 广东, 江门, 529020;
    2. 华南理工大学系统工程研究所, 广东, 广州, 510640
  • 收稿日期:2008-03-14 修回日期:2008-05-14 出版日期:2009-07-20 发布日期:2010-01-03
  • 作者简介:赵良辉(1973- ),男,讲师,博士,主要研究方向为系统工程,人工智能.E-mail:xingzhe7249@163.com
  • 基金资助:
    教育部高校博士学科点专项科研基金(20070561081);广东省工业科技攻关计划(2007B010200046)资助课题

Niche-based immune algorithm for job shop scheduling problems

ZHAO Liang-hui1, YANG Hai-dong2   

  1. 1. School of Management, Wuyi Univ., Jiangmen 529020, China;
    2. Inst. of Systems Science, South China Univ. of Technology, Guangzhou 510640, China
  • Received:2008-03-14 Revised:2008-05-14 Online:2009-07-20 Published:2010-01-03

摘要: 针对免疫算法在解决组合排序问题时一般表现不佳的问题,采用多克隆算子以及独特的浓度控制机制形成具有增强搜索能力的新型免疫算法。多克隆算子与遗传算法中的交叉算子近似,它拓宽了普通免疫算法仅凭高变异方式形成的狭窄搜索空间;基于小生境的浓度控制机制借鉴生物学上的小生境概念,通过相似个体群中选择概率的不均衡分配有效避免算法掉入局部陷阱。所构造的小生境免疫算法在对多个作业车间调度算例的仿真过程中体现了较好的效果。

Abstract: To improve the inefficiency of the immune algorithm in solving combinatorial optimization problems including job shop scheduling problems,this paper adopts the multi-clone operator and a unique density-control scheme to improve the immune algorithm’s optimizing ability.The multi-clone operator is introduced from the genetic algorithm to widen the searching space of the problem,and the density-control scheme is enlightened by the concept "niche" in biology,which can lower similar individual group’s propagation probability among the population to avoid local trap in searching.The algorithm is designed for job shop scheduling problems,whose effect is validated by a series of job-shop scheduling benchmark problems.

中图分类号: