Artificial bee colony algorithm with an adaptive search manner and dimension perturbation

Artificial bee colony (ABC) can effectively solve some complex optimization problems. However, its convergence speed is slow and the exploitation capacity is insufficient at the last search stage. In order to tackle these issues, this paper proposes a modified ABC with an adaptive search manner and dimension perturbation (called ASDABC). There are two important search manners: exploration and exploitation. A suitable search manner is beneficial for the search. An explorative search strategy and another exploitative search strategy are selected to build a strategy pool. To adaptively choose an appropriate search manner, an evaluating indicator is designed to relate the current search status. According to the evaluating indicator, an adaptive method is used to determine which kind of search manner is suitable for the current search. Additionally, a dynamic dimension perturbation strategy is used to enhance the exploration and exploration ability. To verify the performance of ASDABC, 50 problems are tested including 22 classical functions and 28 complex functions. Experiment result shows that ASDABC achieves competitive performance when contrasted with seven different ABC variants.

Automated summary

This paper proposes a modified artificial bee colony algorithm, ASDABC, which performs well in solving complex optimization problems.