Artificial bee colony with bidirectional search (Q2224147)

Summary: Artificial bee colony (ABC) is an efficient meta-heuristic, which has shown good search abilities on many optimisation problems. In the standard ABC, its solution updating equation uses a random weight to control the difference vector between the current food source and another randomly selected one, and the random weight between \(-1\) and 1 determines the search directions. How to select a good search direction is not an easy task. In this paper, we propose a new ABC algorithm called ABC with bidirectional search, which employs a new method to control the search direction of ABC. To verify the effectiveness of the bidirectional search method, we embed it into two ABC variants and test compares their performance with their corresponding parent algorithms. Experimental results on six benchmark functions show that the bidirectional search method can effectively improve the performance of ABC.