Deakin University
Browse

File(s) under permanent embargo

Modification on enhanced Karnik–Mendel algorithm

Version 2 2024-06-04, 02:18
Version 1 2016-12-05, 14:23
journal contribution
posted on 2024-06-04, 02:18 authored by SM Salaken, Abbas KhosraviAbbas Khosravi, S Nahavandi
Karnik–Mendel(KM) algorithm and its enhancements are among the most popular type reduction algorithms in the literature. Enhanced KM (EKM) algorithm is computationally fast and can quickly locate left and right switch points in just a few iteration. This paper proposes a subtle yet very effective modification to EKM algorithm to further improve its computational requirement. The modification relates to how the initial right switch point is determined. Comprehensive simulation results for different cases and scenarios provide the statistical proof for the validity of conclusions drawn on the superiority of the proposed initialization compared to initialization used in the original EKM algorithm. The superiority is quantitatively measured in the number of saved iterations and convergence speed.

History

Journal

Expert Systems with Applications

Volume

65

Pagination

283-291

Location

Amsterdam, The Netherlands

ISSN

0957-4174

eISSN

1873-6793

Language

English

Publication classification

C1 Refereed article in a scholarly journal

Copyright notice

2016, Elsevier Ltd

Publisher

PERGAMON-ELSEVIER SCIENCE LTD