基于交叉算子的改进人工蜂群算法

Modified artificial bee colony algorithm based on crossover operators

针对人工蜂群算法存在后期收敛速度慢、局部搜索能力差和易陷入局部最优的问题,提出一种基于交叉算子的改进人工蜂群算法.该算法利用佳点集方法产生初始种群,使得初始化个体尽可能均匀地分布在搜索空间;随机选择食物源位置与当前最优食物源位置进行算术交叉操作,引导群体向全局最优解靠近,提高算法的局部搜索能力和加快收敛速度.通过5个高维标准测试函数的实验结果表明新算法的有效性.

Aimed at the problems of standard artificial bee colony（ABC）algorithm,such as the low convergence rate,poor local searching ability and easy to be trapped into local optimums,an improved ABC algorithm is proposed based on crossover operator.In this algorithm,an initial colony is generated with optimal point set method to make the initialized individuals distribute as uniformly in the search space as possibly.The arithmetic crossover operation of randomly selected food source position individual and current optimal food source position is carried out,leading the population to approach closely to the global optimum solution,improving the local searching ability,and accelerating the convergence speed.The experimental results of 5 high-dimensional benchmark functions show that the proposed new algorithm is effective.