ESTIMATION OF INFILTRATION PARAMETERS USING GENETIC ALGORITHMS

Abstract

In hydrology infiltration is one of the most important tasks. Accurate determination of infiltration rate is necessary fOr proper determination of surface runoff and groundwater recharge. In this study the priori identifiability and efficacy of Genetic Algorithms in estimating infiltration parameters are discussed. For this purpose two infiltration models (Norton's and Philip's) are chosen. The efficacy of Genetic Algorithms and the • performance of different objective functions are considered. The present study is aimed at evaluating the bias inducted by objective functions. For this purpose, Gaussian noise is added to the synthetically generated infiltration data. The results obtained in this procedure are quite satisfactory for Philip's equation as compared to Horton's equation. Genetic Algorithm is also applied to analyze a set of reported data of Alexis silt loam.