一种简单高效的改进人工蜂群优化算法

摘要

Artificial bee colony algorithm is a novel bio-inspired intelligence optimization algorithm.Compared with other bio-inspired intelligence optimization algorithms,the optimization strategy of artificial bee colony(ABC) algorithm still need to be improved to enhance the convergence speed and the optimization precise.A simple and effective modified artificial bee colony algorithm based on normal distribution is proposed here.Firstly,the nectar source initialization strategy based on normal distribution is given.The purposiveness of the initialization process is improved and the search precise can be ensured.Then,the basic position and the zoom factor in the search equation are modified.The search range is enlarged and the purposiveness of the search is also improved.Therefore,the property of global convergence and the optimization precise are also improved in the proposed modified ABC algorithm.The optimization experimental results for high-dimensional benchmark functions indicate that the proposed modification strategies are simple and effective with better convergence speed and optimization precise.%人工蜂群(Artificial bee colony,ABC)算法是一种新型的仿生智能优化算法.与其他仿生智能优化算法相比,ABC算法的优化求解策略仍有待改进,以进一步提高其收敛速度和优化求解精度.为此,本文提出一种简单而高效的改进ABC算法,将统计学中的正态分布理论引入ABC算法的优化求解过程.首先,提出基于正态分布的蜜源初始化策略,提高了初始化过程的目的性,为后续搜索提供了精度保障.进而对搜索公式中的基础位置和缩放因子进行改进,提出了基于正态分布的搜索策略.该策略在扩大搜索范围的同时,使搜索更新过程更具目的性,从而在有效防止陷入局部收敛的同时,提高了优化求解速度.针对高维复杂Benchmark函数的测试实验结果表明,所提出算法的改进策略简单有效,其收敛速度和求解精度更高.