A Ladder Spherical Evolution Search Algorithm

Haichuan YANG  Shangce GAO  Rong-Long WANG  Yuki TODO  

Publication
IEICE TRANSACTIONS on Information and Systems   Vol.E104-D   No.3   pp.461-464
Publication Date: 2021/03/01
Publicized: 2020/12/02
Online ISSN: 1745-1361
DOI: 10.1587/transinf.2020EDL8102
Type of Manuscript: LETTER
Category: Fundamentals of Information Systems
Keyword: 
spherical evolution,  ladder descent,  population update,  

Full Text: PDF(512.8KB)>>
Buy this Article




Summary: 
In 2019, a completely new algorithm, spherical evolution (SE), was proposed. The brand new search style in SE has been proved to have a strong search capability. In order to take advantage of SE, we propose a novel method called the ladder descent (LD) method to improve the SE' population update strategy and thereafter propose a ladder spherical evolution search (LSE) algorithm. With the number of iterations increasing, the range of parent individuals eligible to produce offspring gradually changes from the entire population to the current optimal individual, thereby enhancing the convergence ability of the algorithm. Experiment results on IEEE CEC2017 benchmark functions indicate the effectiveness of LSE.