Affiliations: [a] School of Electrical and Information Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China | [b] China Nuclear Power Technology Corporation, LTD, Wuhan, Hubei 430223, China | [c] School of Computing, Engineering and Built Environment, Glasgow Caledonian University, Glasgow G4 0BA, UK
Correspondence:
[*]
Corresponding author: Xu Chen, School of Electrical and Information Engineering, Jiangsu University, Zhenjiang, Jiangsu 212013, China. E-mail: xuchen@ujs.edu.cn.
Abstract: Micro differential evolution (mDE) refers to algorithms that evolve with a small population to search for good solutions. Although mDEs are very useful for resource-constrained optimization tasks, the research on mDEs is still limited. In this paper, we propose a new mDE, i.e., vectorized bimodal distribution based mDE (called VB-mDE). The main idea is to employ a vectorized bimodal distribution parameter adjustment mechanism in mDE for performance enhancement. Specifically, in the VB-mDE, two important control parameters, i.e., scale factor F and crossover rate CR, are adjusted by bimodal Cauchy distribution. At the same time, to increase the population diversity, the scale factor F is vectorized. The proposed VB-mDE is evaluated on the CEC2014 benchmark functions and compared with the state-of-the-art mDEs and normal DEs. The results show that the proposed VB-mDE has advantages in terms of solution accuracy and convergence speed.
