Shape Parameter Optimization in Radial Basis Function by Grey Wolf Method

ÇELİK E.

International Journal of Computational Methods, 2026 (SCI-Expanded, Scopus) identifier identifier

  • Publication Type: Article / Article
  • Publication Date: 2026
  • Doi Number: 10.1142/s0219876226500283
  • Journal Name: International Journal of Computational Methods
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Compendex, MathSciNet, zbMATH
  • Keywords: grey wolf optimizer, image zoom-in, multiquadratics, Radial basis function (RBF), shape parameter optimization
  • Çanakkale Onsekiz Mart University Affiliated: Yes

Abstract

This paper presents research on optimizing the shape parameter in radial basis function (RBF) interpolation through the implementation of the grey wolf optimization (GWO) algorithm to enhance efficiency and effectiveness in two-dimensional function interpolation and image zooming task. The methodology involves testing the method on benchmark functions and image datasets, then comparing the outcomes with MATLAB’s GlobalSearch algorithm and the selection of random parameters. Although the RBF consistently demonstrated high-precision capability, this accuracy is predominantly achieved under severe ill-conditioning of the RBF interpolation matrix. Also, the method exhibits improved computational efficiency, performing up to 13 times faster than the GlobalSearch algorithm in image-zooming applications, highlighting the potential of nature-inspired optimization techniques in scientific computing and image processing.