A Bird Flock Gravitational Search Algorithm Based on the Collective Response of Birds. (11th September 2017)

• Record Type: Journal Article
• Authors: Han, XiaoHong; Qiang, Yan; Lan, Yuan
• Is Part Of: Computer journal. Volume 60:Number 11(2017)
• Page Start: 1687
• Page End: 1716
• Publication Date: 2017-09-11
• Subjects: gravitational search algorithm -- learning algorithm -- heuristic search algorithm -- collective behavior -- numerical function optimization -- data clustering; Computers -- Periodicals; 005.1
• DOI: 10.1093/comjnl/bxx048
• Languages: English
• ISSNs: 0010-4620

Abstract

ABSTRACT: Gravitational search algorithm (GSA) is a stochastic search algorithm based on the law of gravity and mass which is widely used nowadays for efficient solution of optimization problems. For the purpose of enhancing the performance of original GSA, this paper proposes a new GSA called Bird Flock Gravitational Search Algorithm (BFGSA) based on the collective response of birds. Although GSA performs well in many problems, algorithms in this category lack mechanisms which add diversity to exploration in the search process. Our proposed algorithm introduces a new mechanism into GSA to add diversity, a mechanism which is inspired by the collective response behavior of birds. This mechanism performs its diversity enhancement through three main major steps including initialization, identification of the nearest neighbors and orientation change. The initialization is to generate candidate populations for the second steps and the orientation change updates the position of objects based on the nearest neighbors. Due to the collective response mechanism, the BFGSA explores a wider range of the search space and thus escapes suboptimal solutions. The efficiency and robustness of the proposed algorithm is demonstrated using multiple traditional and newly composed benchmark functions presented in CEC2005 competition and the results are compared with recent variants of the original particle swarm optimization and state-of-the-art GSA algorithms. Furthermore, we applied BFGSA to a real-world application of data clustering. The results show that BFGSA improves the performance of the original GSA and obtains the best results compared with our selected GSA-type algorithms in benchmarking experiments and clustering experiments.