求解高维复杂优化问题的改进人工蜂群算法

Improved artificial bee algorithm for high dimensional complex optimization problems

为了提高人工蜂群算法求解高维复杂优化问题的能力,提出一种改进人工蜂群算法(artificial bee colony algorithm with attractor,BAABC)。在观察蜂阶段,BAABC算法摒弃轮盘赌选择策略,并通过引进吸引子改变观察蜂的搜索方式。首先,全局最优解波动产生吸引子。然后,观察蜂以吸引子为中心等比例收缩,共同开发同一区域,从而提高了算法的开发能力。实验结果表明,BAABC开发能力显著增强。关于迭代次数和时间,收敛速度都明显提高。在解决高维复杂优化问题方面,BAABC算法优势明显。值得一提的是,BAABC算法的收敛效果与问题维数无关,具有很好的鲁棒性。

To solve the low capacity for high dimensional complex optimization problems, an improved artificial bee colony（BAABC）is proposed in this paper. In the stage of onlooker bees, BAABC algorithm has abandoned Roulette strategy which method onlooker bees selected nectar by, and the search way is changed for onlooker bees by introducing attractor.The attractor is generated by disturbing the global optimal solution. All onlooker bees move toward to the attractor in the same proportion. All onlooker bees together exploit the same area. So the exploitation capacity is enhanced. Experimental results show that the exploitation capacity of global search is remarkably enhanced. About the iteration number or time, convergence speed are improved obviously. To solve high dimensional complex optimization problem, the advantage of BAABC algorithm is obvious. What is more, the convergence performance of BAABC algorithm has nothing to do with the dimension of problem, and the robustness of BAABC is very strong.