A gravitational search algorithm based on levy flight. Issue 4 (April 2021)

• Record Type: Journal Article
• Authors: Zhao, Jing; Zhu, Haidong; Hu, Yinhua; Hu, Enjun; Huang, Baole; Zhang, Tingyu; Zhang, Pan
• Is Part Of: Journal of physics. Volume 1865:Issue 4(2021)
• Publication Date: 2021-04
• Subjects: Gravitational search algorithm -- levy flight -- gravitational constant -- function test; Physics -- Congresses; 530.5
• DOI: 10.1088/1742-6596/1865/4/042006
• Languages: English
• ISSNs: 1742-6588

Abstract

Abstract: A gravitational search algorithm based on levy flight is proposed to solve the problem that the gravitational search algorithm falls into the local optimal solution. Firstly, the dynamic adjustment strategy of gravitational constant is introduced to increase the diversity of particles and improve the convergence speed and accuracy of the algorithm. Then, on the basis of gravitation search algorithm, levy flight is introduced, and the particle search range is increased by using levy flight search mode combining small step and long stride search, so as to improve the global search capability of the algorithm. Finally, six standard test functions are used for simulation analysis. The results show that the improved algorithm has obvious advantages in both convergence speed and convergence precision, and it is more advantageous in multi-peak function optimization.