Automatic calibration of soil–water characteristic curves using genetic algorithms

An automatic technique for the determination of the coefficients of models for soil–water characteristic curves (SWCC) or water retention curves (WRC) is presented. The technique is based on optimisation using genetic algorithms, in which the error between predictions and experimental data is minimised by varying the model parameters. The method is powerful and reasonably efficient in finding the best parameters. Four models are analysed including one accounting for hysteresis behaviour. Details of a simple genetic algorithm (SGA) and its complete application are explained. To account for the hysteresis of the SWCC, the models are programmed in a rate form, in which numerical integration is employed to advance the state variables. One advantage of the optimisation presented is that the best curves averaging both the drying and wetting paths are obtained when hysteresis is present.