自适应贪婪搜索的人工蜂群算法

Abstract

人工蜂群算法是受蜜蜂觅食行为启发提出的一种群体智能优化算法,为了增强人工蜂群算法的开采性能,本文更好地模拟了观察蜂的觅食行为,提出一种自适应贪婪搜索的改进人工蜂群算法,在观察蜂阶段,搜索半径自适应减小,成功搜索某食物源之后可以贪婪地再次搜索该食物源,以充分利用成功的搜索经验,减小搜索盲目性。在10个标准测试函数上的实验表明,改进算法的收敛精度超过ABC和最近提出的q ABC算法,而计算复杂度低于这两种算法。

Artificial bee colony （ABC） algorithm inspired by the foraging behaviour of the honey bees is one of the swarm intelli-gence based optimization techniques. Adaptive greedy search ABC （ AGS-ABC） is a new- version of ABC algorithm in order to enhance the exploitation performance of ABC, which models the behavior of onlooker bees more accurately.In the phase of onlooker bees, the search radius shrinks adaptively and the onlooker bees can search the same food source again after a successful search on the food source in order to make the best of successful search experience and diminish the blind search.Experiments on 10 bench-mark functions show that AGS-ABC outperfor^ms ABC and recently developed quick ABC（qABC） in terms of convergence accuracy and have less complexity compared to the two algorithms.