Abstract: In order to overcome the problems of FOA, such as low convergence precision and unstable convergence resulted from improper random parameter, an improved FOA is proposed, called Fruit Fly Optimization Algorithm with Adaptive Parameter（FOAAP）. In each evolutionary generation, the accurate values describing the characteristics of the overall species are input, 3 digital characteristics[C(Ext,Ent,Het)] of the contemporary cloud model are obtained by backward cloud generator, then using [U]conditions membership cloud generator, the parameter [Value] is adaptively adjusted, which is Fruit Fly’s searching distance and direction for food. FOAAP is compared with FOA and other algorithms in reference literatures, experimental results show that FOAAP has the advantages of speeder convergence, higher convergence precision and higher convergence reliability.

Key words: cloud model, adaptive, Fruit Fly Optimization Algorithm（FOA）, convergence precision

摘要： 针对基本果蝇优化算法因参数选取不当而导致的收敛精度偏低且不稳定的问题，提出了自适应调整参数的果蝇优化算法（FOA with Adaptive Parameter，FOAAP）。该算法在每个进化代输入描述种群整体特征的精确数值，由逆向云发生器算法得到当代云模型的3个数字特征[C(Ext,Ent,Het)]，按照[U]条件隶属云发生器自适应调整果蝇个体搜寻食物的方向与距离[Value]这一参数。将该算法在函数优化中，与基本果蝇优化算法以及相关文献中算法进行仿真对比，结果表明，新算法在收敛速度、收敛可靠性及收敛精度方面具有明显优势。

关键词: 云模型, 自适应, 果蝇优化算法, 收敛精度