摘要
为了改善排课的收敛性与效率,提出了一种基于多种群遗传算法的排课方法。在算法中根据杂种优势理论的原理,让多个种群同时进化,改变了传统的遗传算法在单个种群中演化繁衍。多种群之间既竞争又合作,共同寻找全局最优解,提高了算法的收敛速度。该算法摒弃了完全随机搜索的做法,依据适应度函数中各项权重比例的多寡为导向,定向随机生成染色体中的基因,从而提高了算法的效率。最后,通过两组实验数据表明了该算法的收敛性与高效率。
In order to improve the convergence and efficiency of course arrangement, a new algorithm based on multi-population genetic algorithm is presented, which alters the traditional genetic algorithm of breeding in a single population, and instead evolving in multi- population simultaneously according to heterosis theory. They both compete and collaborate so as to seek for global optimal solution to- gether, which increases the astringency of algorithm. It abandons the behavior of searching targets in a complete random way, and instead depends on the amount of the weight percentage of the fitness function to generate the genes in the chromosome in a directional random way, thus it enhances the efficiency of algorithm. Lastly, the convergence and high efficiency of the new algorithm are verified with two groups of data obtained from experiments.
出处
《计算机工程与设计》
CSCD
北大核心
2010年第22期4877-4880,4908,共5页
Computer Engineering and Design
关键词
多种群
配子
排课
定向随机搜索
跳跃基因
杂种优势
multi-population
gamete
course arrangement
directional randomsearch
jumping gene
heterosis