A competitive mechanism based multi-objective differential evolution algorithm and its application in feature selection

作者:

Highlights:

• An efficient competitive mechanism based multi-objective differential evolution algorithm (CMODE) is proposed.

• Two variants of the CMODE using the angle competitive mechanism and Euclidean distance competitive mechanism are proposed.

• The performance of our proposed CMODE is comprehensively assessed by comparing with six popular MOEAs on twenty benchmarks.

• The proposed CMODE is applied to the feature selection problem.

摘要

•An efficient competitive mechanism based multi-objective differential evolution algorithm (CMODE) is proposed.•Two variants of the CMODE using the angle competitive mechanism and Euclidean distance competitive mechanism are proposed.•The performance of our proposed CMODE is comprehensively assessed by comparing with six popular MOEAs on twenty benchmarks.•The proposed CMODE is applied to the feature selection problem.

论文关键词:multi-objective algorithm,Differential evolution,Competitive mechanism,Feature selection

论文评审过程:Received 17 April 2021, Revised 21 February 2022, Accepted 9 March 2022, Available online 17 March 2022, Version of Record 31 March 2022.

论文官网地址:https://doi.org/10.1016/j.knosys.2022.108582