A near-surface interpolation scheme based on radial basis function

The interpolation points adopted in the radial basis function (RBF) can be scattered. Based on the above fact, a near-surface interpolation scheme is proposed to combine RBFs for the scattering problems modeled with surface integral equations. The interpolation efficiencies of different RBFs with the proposed and Tartan grid schemes are compared to approximate the interactions between well-separated groups. It can be seen from the numerical results that the number of interpolation points is reduced significantly for all four RBFs, and the accuracy of the Gaussian RBF is better than the other three RBFs for different sizes of groups. The proposed scheme with GA RBF can be employed to build a fast solver.