Automatic Fit of the Variogram

Abstract

The automatic fit of the experimental variogram is a difficult problem of geostatistics. In order to solve the fitting problem, an improved linear programming method is presented in this paper. In combination with the advantages of weighted polynomial method and linear programming method, the new method uses the inverse of lag distance as weight coefficient. We can rationally estimate the parameters of a spherical model by this method. For illustration, an example is utilized to show the feasibility of the improved method in solving the fitting problem. Empirical results show that the fittd curve of spherical theoretical model, which is obtained by this method, is more representative of the experimental variogram values. The improved linear programming method can effectively solve the automatic fitting problem of the experimental variogram, and a uniform and objective estimation of optimal parameters will be obtained by this method.