Abstract:

As the size of the optimization problems grows, the solutions are increasingly difficult. Many intelligent algorithms can be used to overcome these problems. One of the many intelligent algorithms is the Artificial Bee Colony (ABC) algorithm. Some improvements in the original ABC algorithm are needed to take advantage of the artificial bee colony algorithm in large-scale optimization problems. In this study, the improvements to the ABC algorithm are defined in a new ABC algorithm we call "Elite-based Self-Adjustable Artificial Bee Colony algorithm". Unlike the classic ABC algorithms, different search equations were used in the workers and observers’ bees steps, and these search equations were used by elite individuals. Also, the performance of the algorithm has been enhanced with a local search technique. The correct selection of the parameters of the algorithms has a great impact on the success of the algorithms. In this study, the Irakian tool has been used to best adjust the parameters of the algorithm. The algorithm we have developed has been tested in the SOCO11 measurement set of functions that contains large-scale continuous optimization functions. The results we obtained were compared with the ABC algorithms and the algorithms involved in the SOCO11 competition and successful results were obtained.

Keywords: