关键词: 人工蜂群算法, 禁忌搜索, 局部最优

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