基于Hooke-Jeeves的改进人工蜂群算法

An Improved Artificial Bee Colony Algorithm Based on Hooke-Jeeves Method

人工蜂群算法(ABC)是新近提出的一种基于群智能的优化方法,它比其他基于种群的智能算法更优异,但蜂群的搜索更新公式在算法的局部寻优能力上存在缺陷。因此本文致力于将擅长局部寻优的搜索机制引入人工蜂群算法,提出了一种基于Hooke-Jeeves的改进人工蜂群算法(IHABC)。IHABC算法改进了采蜜蜂和跟随蜂的搜索公式,期望保留全局搜索能力的同时能更大限度地增加算法的局部寻优能力;采用质量中上的个体优化Hooke-Jeeves搜索法的初始基点以确保局部寻优的有效性;改进Hooke-Jeeves方法的探索移动的步长公式。为了检测新算法的性能,将其与人工蜂群算法、Hooke-Jeeves人工蜂群算法(HABC)进行比较分析,30个基准函数上的数值实验结果表明,IHABC算法在求解无约束优化问题时得到的近似解有更高的精度。

Artificial bee colony algorithm (ABC) is a relatively new swarm intelligence optimization method, which is superior to other population-based intelligent algorithms. However, ABC algorithm also has certain limitation because its updating formula is not good at exploitation. In order to enhance the exploitation capacity of ABC, this paper presents a new algorithm called an improved artificial bee colony algorithm based on Hooke-Jeeves method (IHABC). The altered formulas of employed bees and onlooker bees in IHABC not only keep exploration ability but also increase exploitation to a great extent. In addition, the algorithm optimizes initial base point selection in Hooke-Jeeves search phase by upper-middle individual and modifies step size formula of exploratory move, so that the whole population evolves spontaneously in the right direction. To test the effectiveness of the proposed algorithm, we compare IHABC with ABC and Hooke-Jeeves artificial bee colony algorithm (HABC). The numerical experimental results of 30 benchmark functions clearly indicate that IHABC gets higher approximate solution precision in solving unconstrained optimization problems.