An improved multi-objective evolutionary algorithm based on point of reference

2017 International Symposium on Application of Materials Science and Energy Materials

材料科学;能源学

Zhang, Boyi^1 ; Zhou, Xue^1 ; Liu, Yuqing^1 ; Xu, Xiangli^2 ; Zhang, Libiao^1

School of Information Science and Technology, Northeast Normal University, Changchun, China^1

Department of Information Operation and Command Training, NDU of PLA, Beijing, China^2

关键词: Crowding distance;    Multi objective evolutionary algorithms;    Nondominated solutions;    Reference points;    Test functions;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/322/6/062025/pdf
DOI  :  10.1088/1757-899X/322/6/062025

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

In the article presents a new evolutionary algorithms, this algorithm is based on reference points. This algorithm according to a archive find the non-dominated solution, and make this solution from large to small order according to crowding distance and let the space is divided to several small space equidistant, in each the subspace to obtain its centroid as the reference point. The method used in this paper are not predefined reference point but in the course of evolution according to the current state of dynamic reference point particles to guide the evolution, finally using five effectiveness test functions prove algorithm.