A quasi-RBF technique for numerical discretization of PDE's

Despite such very appealing features of the radial basis function (RBF) as inherent meshfree and independent of dimension and geometry, the various RBF-based schemes [1] of solving partial differential equations (PDE’s) still confront some deficiencies. For instances, the lack of easy-to-use spectral convergent RBFs, ill-conditioning and costly evaluation of full interpolation matrix. The purpose of this study is a combined use of radial basis function and the other approximation methods to cure these drawbacks.