Abstract: To solve the shortcoming of neighborhood search and the local optimum of Artificial Bee Colony algorithm（ABC）, this paper imports the Taboo Search（TS）, and proposes the ABC based on Taboo Search（TS-ABC）. This paper imports two taboo tables: T1 and T2. The length of T1 is finite. T1 store the present solutions which the bees visit. The length of T2 is infinite. T2 stores the solutions which can’t be better after limit times of visiting. When the bees search new solution near honeymoon’s location, they will jump the solution in the T1 or T2. This kind of design avoids the repeating search, enhances the ability of neighborhood searching and overcomes local optimum. Experiments on 15 benchmark functions show that the performance of TS-ABC is priori to ABC；the performance of TS_ABC is more priori to ABC solving multimodal functions; with the increase of dimension, the performance of TS_ABC is better than ABC.

Key words: Artificial Bee Colony algorithm, taboo search, local optimum

摘要： 针对人工蜂群算法（Artificial Bee Colony，ABC）邻域搜索能力不强且容易陷入局部最优的不足，引入禁忌搜索的思想，提出了基于禁忌搜索的人工蜂群算法（TS_ABC）。TS_ABC算法在ABC算法的基础上加入两个禁忌表，分别记为禁忌表T1和禁忌表T2。禁忌表T1的长度是有限的，存储蜜蜂访问过的当前解；禁忌表T2的长度是无限的，存储优化[limit]次后没有改进的解。蜜蜂在蜜源位置搜索新解时要跳过禁忌表里的解，这样避免了重复搜索，增强了邻域搜索能力，克服了容易陷入局部最优。15个标准函数上实验结果表明：（1）TS_ABC的性能优于ABC算法；（2）在求解多峰函数最优解时，TS_ABC性能更加优于ABC算法；（3）随着函数维数的增加，相对于ABC算法，TS_ABC性能提高更多。3个标准函数上实验结果表明：TS_ABC算法性能优于ABC算法，即提出的使用两个禁忌表的方法优于只使用一个禁忌表的方法。

关键词: 人工蜂群算法, 禁忌搜索, 局部最优