Publication Type : Conference Proceedings
Publisher : Pattern Recognition and Machine Intelligence, Springer International Publishing
Source : Pattern Recognition and Machine Intelligence, Springer International Publishing, Volume 11941, Cham, p.328-336 (2019)
Url : https://link.springer.com/chapter/10.1007/978-3-030-34869-4_36
ISBN : 9783030348694
Campus : Coimbatore
School : School of Engineering
Department : Computer Science
Year : 2019
Abstract : Differential Evolution (DE) is a popular optimization algorithm in the repository of Evolutionary Algorithm (EAs). The DE algorithm is known for its simple algorithmic structure, which has minimal number (only three) of control parameters. A propitious avenue for enhancement of DE’s performance is making it a self-adaptive algorithm. There exist many algorithms for self-adapting one or more of DE parameters. The self-adaptiveness of any parameter needs critical analysis on the impact of that parameter. This paper analyzes and presents the impact of the parameter - mutation scale factor (F) of DE. Including empirical evidences for understanding the effect of F on the nature of convergence of DE at solving a problem is the novelty of this paper. The experiment includes implementing a set of benchmark functions, with diversified features, using different variants of DE, in order to critically analyze the role of F.
Cite this Research Publication : Dhanya M. Dhanalakshmy, Dr. Jeyakumar G., and C. Velayutham, S., “Analytical Study and Empirical Validations on the Impact of Scale Factor Parameter of Differential Evolution Algorithm”, Pattern Recognition and Machine Intelligence, vol. 11941. Springer International Publishing, Cham, pp. 328-336, 2019.