A competitive swarm optimizer integrated with Cauchy and Gaussian mutation for large scale optimization
Qiang Zhang, Hui Cheng, Zhencheng Ye, Zhenlei Wang · 2017
Competitive swarm optimizer (CSO) has shown promising results for solving large scale global optimization problems proposed recently. However, CSO shows insufficient exploitation of the population. In this paper, a competitive swarm optimizer integrated with Cauchy and Gaussian mutation (CGCSO) is proposed for large scale optimization. The new algorithm does not only update the losers' positions with the CSO method, but also update the winners' positions by Cauchy and Gaussian mutation to improve the exploitation capability of the population. Moreover, CGCSO utilizes the ring topology to enhance the swarm diversity and alleviate premature convergence. A comparative study between CGCSO and CSO evaluated on the CEC'08 benchmark functions has been carried out. The experimental results indicate that CGCSO performs better on the whole, especially on the non-separable functions and 500-dimensional problems.