An improved gravitational search algorithm combining with centripetal force

Abstract

Although GSA has strong search ability, it is still easy to fall into local optimum and accuracy is often not high enough. Inspired by the concept of centripetal force, we propose a centripetal force based gravity search algorithm, named CF-GSA. It controls the parameters in different stages, which balances the exploration and exploitation capabilities of particles. We performed comparative experiments on 23 benchmark functions, and analyzed the performance of this algorithm. In experiments, it can effectively prevent falling into local optimun and achieve higher accuracy and better convergence.

Keywords

• Gravity search algorithm
• Centripetal force
• Meta-heuristic
• Convergence