location: Current position: Home >> Scientific Research >> Paper Publications

Hybrid Artificial Bee Colony Algorithm with Differential Evolution and Free Search for Numerical Function Optimization

Hits:

Indexed by:期刊论文

Date of Publication:2016-08-01

Journal:INTERNATIONAL JOURNAL ON ARTIFICIAL INTELLIGENCE TOOLS

Included Journals:SCIE、EI、Scopus

Volume:25

Issue:4

ISSN No.:0218-2130

Key Words:Artificial bee colony algorithm; differential evolution; free search selection; chaotic systems; numerical function optimization

Abstract:Artificial bee colony (ABC) algorithm invented by Karaboga has been proved to be an efficient technique compared with other biological-inspired algorithms for solving numerical optimization problems. Unfortunately, convergence speed of ABC is slow when working with certain optimization problems and some complex multimodal problems. Aiming at the shortcomings, a hybrid artificial bee colony algorithm is proposed in this paper. In the hybrid ABC, an improved search operator learned from Differential Evolution (DE) is applied to enhance search process, and a not-so-good solutions selection strategy inspired by free search algorithm (FS) is introduced to avoid local optimum. Especially, a reverse selection strategy is also employed to do improvement in onlooker bee phase. In addition, chaotic systems based on the tent map are executed in population initialization and scout bee's phase. The proposed algorithm is conducted on a set of 40 optimization test functions with different mathematical characteristics. The numerical results of the data analysis, statistical analysis, robustness analysis and the comparisons with other state-of-the-art-algorithms demonstrate that the proposed hybrid ABC algorithm provides excellent convergence and global search ability.

Pre One:Integrated post-disaster medical assistance team scheduling and relief supply distribution

Next One:人工蜂群算法在单时间窗车辆路径问题中的应用