An adaptive artificial bee colony algorithm based on objective function value information

Xiaoyu Song

doi:10.1016/j.asoc.2017.01.031

2017-06-01

PRE

Xiaoyu Song

Qifeng Yan *

Ming Zhao

DOI：10.1016/j.asoc.2017.01.031

原文链接

原文求助

摘要

En 中文

Artificial bee colony (ABC) algorithm is a novel biological-inspired optimization algorithm, which has many advantages compared with other optimization algorithm, such as less control parameters, great global optimization ability and easy to carry out. It has proven to be more effective than some evolutionary algorithms (EAs), particle swarm optimization (PSO) and differential evolution (DE) when testing on both benchmark functions and real issues. ABC, however, its solution search equation is poor at exploitation. For overcoming this insufficiency, two new solution search equations are proposed in this paper. They apply random solutions to take the place of the current solution as base vector in order to get more useful information. Exploitation is further improved on the basis of enhancing exploration by utilizing the information of the current best solution. In addition, the information of objective function value is introduced, which makes it possible to adjust the step-size adaptively. Owing to their respective characteristics, the new solution search equations are combined to construct an adaptive algorithm called MTABC. The methods our proposed balance the exploration and exploitation of ABC without forcing severe extra overhead in respect of function evaluations. The performance of the MTABC algorithm is extensively judged on a set of 20 basic functions and a set of 10 shifted or rotated functions, and is compared favorably with other improved ABCs and several state-of-the-art algorithms. The experimental results show that the proposed algorithm has a higher convergence speed and better search ability for almost all functions. (C) 2017 Elsevier B.V. All rights reserved.

Keyword:

Optimization algorithm

Artificial bee colony (ABC) algorithm

Adjust dynamically

Adaptive algorithm

Search equations