A hybrid gravitational search algorithm for unconstrained problems
Xin Zhang, Dexuan Zou, Zhi Kong, Xin Shen · 2018
The basic gravitational search algorithm (GSA) could fall into local optima solution easily, and thus we proposed a hybrid gravitational search algorithm (HGSA) in order to overcome the shortcoming of GSA. This improved gravitational search algorithm, which only uses the position update formula that was affected by the Gbest in its iteration process, is combined with the Differential Evolution (DE) algorithm. Ten benchmark functions have been introduced for testing the improved algorithms' performance. We also use the statistical method "T-test" to verify the difference in results. Experimental results show that HGSA is superior to the basic gravitational search algorithm and its three improved algorithms in terms of both convergence accuracy and convergence rate.