Enhanced Gaining-Sharing Knowledge-based algorithm

Abstract

This article suggests an Enhanced Gaining-Sharing Knowledge-based algorithm (eGSK) to resolve unrestricted optimization minimization problems over a continuous space. This algorithm is based on the idea that people learn and share knowledge throughout their lives. The modification is fundamentally inspired by the principles of Adjust Selection Criteria, Modify Parameters Setup, and Escaping from Local Minimum Solutions, respectively. We conducted comparisons and statistical tests with the Gaining-Sharing Knowledge-based algorithm (GSK) and other algorithms to verify and analyze the performance of the eGSK algorithm. We also performed numerical experiments on 29 test problem sets in 10, 30, 50, and 100 dimensions from the Congress on Evolutionary Computation (CEC) 2017 benchmark. The results were compared with three GSK variant algorithms, seven state-of-the-art algorithms, and GSK alongside components of the eGSK algorithm. According to test results, the eGSK algorithm performs exceptionally well at solving optimization problems with 30, 50, and 100 dimensions and is competitive in 10 dimensions. This means the proposed eGSK algorithm outperforms its competitors and achieves more competitive results, especially with high dimensions.