Genetic Algorithm Based Parameter Estimation of Nash Model

Unit hydrograph method is usually used to resolve surface runoff concentration process. Empirical unit hydrograph is sometimes jagged and not smooth because of estimation errors. Owing to the randomicity of correction method, there is some localization in its application. The application of instantaneous unit hydrograph is relatively wider. Generally, moment method is used to estimate the parameters of instantaneous unit hydrograph. However, the error is obvious between the observed flood process and the predicted flood process with moment method, especially near the flood peak. The genetic algorithm toolbox of matlab software is used to optimize the parameters of instantaneous unit hydrograph. The statistical function gamcdf(x, α, β) in matlab toolbox is used to calculate S(t) curve, which can avoid the errors caused by approximate formula method. The case study indicates that the weighted sum of absolute error applying the method in this paper is 25, the result applying moment method is 63, and the result of approximate formula method is 49. The results show that the method in this paper is more effectual than the other two methods.