摘要
原有蜜蜂交配算法杂交信息量小、勘探能力不足、蜂群多样性少,为了克服这些缺点,对蜜蜂交配算法进行了改进,主要包括交换父代、母代染色体中相互冲突的课程基因增大算法的交叉信息量,设定多种邻域并集的局部搜索策略扩大搜索空间,采用基于k对肯配链的变异操作和雄蜂的部分替换策略丰富蜂群中染色体的组成。应用苏哈数据集进行了测试,测试结果表明,该改进算法较原有算法具有更好的收敛精度、更快的收敛速度,在满足多重约束条件下,能够更有效地解决排课问题。
The original honey-bee mating optimization （HBMO） algorithm has some disadvantages of the crossover operator and the insufficient of exploration while also enlarge the search space for finding the optimal solution. In order to improve the original HBMO, several aspects are adapted including conflicts-based crossover operator, k-pair kempe chain mutation operator, double- neighborhood hill climb, the replacement strategy of drone. A lot of experiments on datasets shown that the improved algorithm increases the convergence precision and accelerates the convergence speed. The course timetabling problems are solved very well in complicated constraint conditions.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第7期2431-2435,共5页
Computer Engineering and Design
基金
天津市应用基础与前沿技术研究计划重点基金项目(11JCZDJC15700)
关键词
蜜蜂交配算法
排课
课程争斗
K对肯配链
双邻域爬山法
HBMO
course arrangement
conflicts-based crossover
k-pair kempe chain
double-neighborhood hill climb
