摘要
本文提出了一种蜜蜂进化型遗传算法.在该算法中,种群的最优个体作为蜂王与被选的每个个体(雄蜂)以概率进行交叉操作,增强了对种群最优个体所包含信息的开采能力.为了避免算法过早收敛,在代进化过程中引入了一个随机种群,提高了算法的勘探能力.通过将该算法建模为齐次有限M arkov链,证明了它的全局收敛性.实验结果表明,蜜蜂进化型遗传算法是一种提高遗传算法性能的有效改进算法.
This paper proposes a Bee Evolutionary Genetic Algorithm （BEGA）. In BEGA, optimum individual being a queen-bee in population crossover with each selected individual （drone）. As a result,it reinforces the exploitation of genetic algorithm. In order to avoid premature convergence, BEGA introduces a random population that extends search area. Consequentially it enhances the exploration of genetic algorithm. By treating the collection of individuals in each generation as a state and modeling the algorithm as a homogeneous finite Markov chain, it is proven that BEGA can guarantee the convergence towards the global optimum of the problem. Experiments results show BEGA is an efficient and effective improved genetic algorithm.
出处
《电子学报》
EI
CAS
CSCD
北大核心
2006年第7期1294-1300,共7页
Acta Electronica Sinica
基金
国家863高科技发展计划基金(No.2001AA422270)
国家自然科学基金(No.69985002)
