[期刊论文]


An Empirical Performance Analysis of Differential Evolution Variants on Unconstrained Global Optimization Problems

作   者:
G. Jeyakumar;C. Shunmuga Velayutham;

出版年:2010

页     码:77 - 86
出版社:Machine Intelligence Research Labs (MIR Labs)


摘   要:

In this paper we present an empirical, comparativeperformance, analysis of fourteen Differential Evolution(DE) variants on different classes of unconstrained globaloptimization benchmark problems. This analysis has beenundertaken, with an objective, to compare and to identifycompetitive DE variants which perform reasonably well onproblems with different features. Towards this, fourteenvariants of DE are benchmarked on 14 high dimensionalunconstrained test functions grouped by their modality anddecomposability viz. unimodal separable, unimodalnonseparable, multimodal separable and multimodalnonseparable. The analysis identifies the overall competitivevariants as well as the feature based performances of all thevariants.



关键字:

Differential Evolution (DE); benchmarking DE;variants; empirical performance analysis; unconstrained numerical;optimization; success performance


全文
所属期刊
International Journal of Computer Information Systems and Industrial Management Applications
ISSN: 2150-7988
来自:Machine Intelligence Research Labs (MIR Labs)