Open Access

ARTICLE

Radial Basis Function and Genetic Algorithms for Parameter Identification to Some Groundwater Flow Problems

by B. Amaziane¹, A. Naji², D. Ouazar³

1 Université de Pau, LMA-FRE2570 CNRS, IPRA, Avenue del’Université, 64000 Pau, France. brahim.amaziane@univ-pau.fr
2 Faculté des Sciences et Techniques, Université Abdelmalek Essaadi, MASE, B. P. 416, Tangier, Morocco. naji@fstt.ac.ma
3 Ecole Mohammadia d’Ingénieurs, Université Mohammed V, LASH, B. P. 765, Agdal, Rabat, Morocco. ouazar@emi.ac.ma

Computers, Materials & Continua 2004, 1(2), 117-128. https://doi.org/10.3970/cmc.2004.001.117

Abstract

In this paper, a meshless method based on Radial Basis Functions (RBF) is coupled with genetic algorithms for parameter identification to some selected groundwater flow applications. The treated examples are generated by the diffusion equation with some specific boundary conditions describing the groundwater fluctuation in a leaky confined aquifer system near open tidal water. To select the best radial function interpolation and show the powerful of the method in comparison to domain based discretization methods, Multiquadric (MQ), Thin-Plate Spline (TPS) and Conical type functions are investigated and compared to finite difference results or analytical one. Through two sample problems in groundwater flow, we demonstrate the computational capacity of RBF in simulating time dependent problems and the possibility of simultaneous estimation of multiple groundwater parameters computational feasibility when it is coupled to simple Genetic Algorithms (GAs). Performance of variously based RBF is compared.

---

Keywords

---