A Modified Artificial Bee Colony Algorithm Using Accept–Reject Method: Theory and Application in Virtual Power Plant Planning

Abstract

In this paper, the modified artificial bee colony optimization algorithm, in which the accept–reject method is used as a part of the solution process in order to limit the random search space, is proposed for solving the optimization problems. The accept–reject method can find points of desirable quality and high accuracy with the help of probabilistic functions and limited the random search space, which increases the speed of the algorithm for the next steps and also reduces the number of iterations of the artificial bee algorithm. Then, the next phases of the bee algorithm will be executed in the interval determined by the accept–reject method. One of the most important issues in intelligent algorithms, especially the bee algorithm, is that in problems with several local optimal points, sometimes the algorithm gets stuck in the trap of local optimal points and does not reach the global optimal points. By using the motion step determination parameter in the employed bee phase, the algorithm can be guided to reach the global optimal points while correctly adjusting the movement of the bees for searching in random food sources, which will help improving the problem-solving process. This algorithm is tested by the basic benchmark functions as well as the benchmark functions of CEC 2005 and CEC2017 and CEC2019 and an engineering problem related to the planning of a virtual power plant. The results show that the proposed algorithm is superior in comparison to other algorithms.