基于轮盘赌反向选择机制的蜂群优化算法

针对人工蜂群算法易陷入局部最优的不足,考虑到基本蜂群算法中个体选择大多基于贪婪选择的思想,从而使算法快速向适应度值高的个体进化而陷入局部停滞。为此,提出一种基于轮盘赌的反向选择机制,以保持蜂群个体的多样性而使算法保持较好进化能力。通过对经典测试函数的仿真实验表明,改进的蜂群算法有更快的收敛速度和更好的收敛精度,且改进的蜂群算法对群体规模有很强的鲁棒性。

Artificial bee colony based on reverse selection of roulette

Towards the defect which inclined artificial bee colonyABC to fall into local minima, and taking into account of greedy selection scheme was always employed during the basic artificial bee colony, which resulted in rapid evolving toward the more fitter individual, and thus trapped the ABC into stagnation. Thus, this paper proposed a modified artificial bee colony MABC based on reverse selection of roulette which retaining the diversity of population in order to improve the evolving capability. Experiments result on a few of benchmark functions show that the MABC algorithm not only effectively avoids the premature convergence, but also significantly improves the convergence speed and the convergence precision. Moreover, the MABC algorithm is robust to the scale of population.