Multi-population kidney-inspired algorithm with migration policy selections for feature selection problems

Najmeh Sadat Jaddi and Salwani Abdullah and Say leng goh and Mohd Zakree Ahmad Nazri and Zalinda Othman and Mohammad Kamrul Hasan and Fatemeh Alvankarian (2025) Multi-population kidney-inspired algorithm with migration policy selections for feature selection problems. IEEE Access., XX. pp. 1-17.

[img] Text
FULL TEXT.pdf
Restricted to Registered users only

Download (1MB) | Request a copy

Abstract

Optimization algorithms often encounter challenges in effectively managing the trade-off between exploration and exploitation, usually leading to less-than-optimal outcomes. This study introduces two novel migration policies in multi-population version of kidney-inspired algorithm (KA) to address this dilemma. The initial algorithm, coded as MultiPop-KA, implements a predetermined migration policy. Conversely, the second algorithm, coded as AutoMultiPop-KA, adopts an adaptive migration policy selection process that determines migration type based on the average fitness of subpopulations. By capitalizing on a multi-population framework and incorporating two migration policies, these methods aim to achieve a more refined equilibrium between exploration and exploitation, thereby augmenting the effectiveness of the KA. Experimental evaluations, conducted across 25 test functions and applied to 18 benchmark feature selection problems, demonstrate the efficacy of the proposed techniques. These results indicate that the proposed approach can significantly enhance optimization algorithms' performance and overall quality.

Item Type: Article
Keyword: Exploration and exploitation; Kidney-inspired algorithm; Multi-population; Migration policy; Feature selection.
Subjects: Q Science > QA Mathematics > QA1-939 Mathematics
Q Science > QA Mathematics > QA1-939 Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Department: FACULTY > Faculty of Computing and Informatics
Depositing User: ABDULLAH BIN SABUDIN -
Date Deposited: 23 May 2025 15:56
Last Modified: 23 May 2025 15:56
URI: https://eprints.ums.edu.my/id/eprint/43874

Actions (login required)

View Item View Item