An Improvement of Gravitational Search Algorithm

Abstract

The gravitational search algorithm (GSA) has the advantages of strong exploitation performance and fast convergence speed. However, the GSA is easy to appear premature phenomenon and get into the local optimum during the search process because the particle diversity declines during the optimization process and the particle swarm optimization information is not shared. Therefore, an improved GSA (IGSA) is proposed, which keeps the particle diversity by adjusting the gravitational constant and enhances particle swarm information sharing ability by imposing global optimal information into the search position of each particle. The proposed IGSA has been evaluated on 9 nonlinear benchmark functions and compared with standard GSA and particle swarm optimization (PSO). The obtained results confirm that the convergence accuracy of IGSA is several orders of magnitude higher than that of GSA, and the search speed is also increased by more than 2 times. In addition, a case study of optimizing generator operating costs of IEEE9 is carried, and the IGSA algorithm achieves lower operating costs and less network loss than that of GSA.