DEVELOPMENT OF GENETIC ALGORITHM FOR IN VERSION OF D.C. RESISTIVITY SOUNDING DATA

Abstract

Implementing Simple Genetic Algorithms (SGAs) as global optimization technique, a MATLAB program is developed for interpretational of one dimension DC resistivity sounding data from Schlumberger configuration of electrode. This program can also be used for any optimization problem by changing the response calculations. In this work, forward response for Schlumberger configuration of electrode is calculated by using Ghosh's inverse filter method. To establish the applicability of the developed program, this is applied for inversion of synthetic VES data, with and without noise, for different layer models. To define model space (or search space), two approaches, 'broad search range method' and 'joint inversion scheme', are discussed. These have been tested on synthetic data and published field data. The results of field data are correlated with well data given in paper. Comparative study of these two approaches shows that the joint inversion scheme is better than broad search range method. This is due to reduction in the computational time and improving efficiency of program. Finally, a field data, comprising five stations on a profile, from the region of Sonbhadra, Uttar Pradesh, India, is inverted and profile section with different depth are plotted. The inverted field results are correlated with the results of paper (Ndatuwong and Yadav, 2013) and stratigraphy and then interpreted geologically