【摘要】 针对人工蜂群算法收敛速度慢以及蜜源的选择概率区分度不够等缺点,提出了一种新的改进人工蜂群算法。依据当前最优蜜源、最差蜜源、当前蜜源建立正弦选择概率模型,并结合全局最优引导策略,构成新算法。概率模型以最优、最差蜜源适应值之差为尺度,衡量当前蜜源适应值所占比重,随后将比重值带入sin函数,即可得当前蜜源的选择概率值。在30、100维上,22个基准测试函数的仿真实验结果表明,正弦选择概率模型能克服后期蜜源区分度不够的问题,为观察蜂跟随雇佣蜂提供正确的引导。与知名的改进人工蜂群算法比较,该算法具有很好的优势。更多还原

【Abstract】 A new improved artificial bee colony algorithm is proposed for the disadvantages of slow convergence speed and inadequate distinguishing of nectar source selection probability. The new algorithm establishes the sinusoidal selection probability model based on the current optimal source,the worst source and the current source of honey,and combines the global optimal guidance strategy. The probability model takes the difference between the optimal and the worst adaptation values as the scale to measure the proportion of the current function values of the honey source,and then the proportion values are substituted into the sin function to obtain the selection probability values of the current honey source. To verify the effectiveness of the new algorithm,based on 22 benchmark functions,this paper compares the algorithm with 6 mainstream improved artificial bee colonies in 30 and 100 dimensions. Experimental results show that the sinusoidal selection probability model can overcome the problem of insufficient nectar source discrimination and provide correct guidance for onlookerbees following Employbees. Compared with other improved artificial bee colony algorithm,this algorithm has unique advantages.更多还原