【作者】 陈知美； 顾幸生；

【Author】 Chen Zhimei, Gu Xingsheng (Research Institute of Automation .East China University of Science and Technology, Shanghai, 200237)

【机构】 华东理工大学自动化研究所；

【摘要】 蚁群算法是一种新型的模拟进化算法,初步的研究表明该算法具有许多优良的性质。为了能应用于解决生产管理和组合优化领域的重要研究方向--Job Shop问题,本文对蚁群算法进行了适当的改进:一是使用易操作的可选工作集,省去了蚁群算法中的禁忌表;二是使用简洁的选择策略,并将总解集分解成与工件数相等的子集。通过实例计算,得到了良好的结果并找到了多种较好的排序方案,同时指出了蚁群算法中还需解决的问题和方向。更多还原

【Abstract】 Ant colony system is a novel simulated evolutionary algorithm, which shows many excellent characters. In this paper, a kind of ant colony system for minimizing the makespan in Job Shop scheduling problem which is an important subject in the fields of production management and combinatorial optimization is presented. The new algorithm includes two important aspects: the tabu list is omitted by using the easily-operational allowed collection, on the other hand, a brief transition strategy with random selection is used and subsets with different initial nodes are separated from the set of solutions. Computational results on some 6 jobs and 6 machines problems show that the proposed algorithm can find optimum solutions. It also points out that the ant colony algorithm still has to be modified in many ways.更多还原