Development of reservoir operation policies considering variable agricultural water demands

摘要

In this paper, a genetic algorithm (GA) optimization model is developed for reservoir operation optimization considering variations in water demands. In order to incorporate the demand uncertainties in optimal operation policies, different types of linear equations with different combinations of inflow, storage at the beginning of the month, and water demands as independent variables have been considered. The coefficients of optimal operation policies are obtained using classic and fuzzy regression analysis. In the case of fuzzy regression, both symmetric and asymmetric membership functions are used. Efficiency of operation policies are evaluated based on the long-term operation simulation of Zayandeh-Rud Reservoir in central part of Iran. Estimated figures for the four criteria of reliability, resiliency, total vulnerability, and maximum monthly vulnerability and also the statistical criteria of correlation coefficient, coefficient of efficiency, and standard error indicate that the fuzzy linear regression equations with inflow, storage, and demand as independent variables, in which asymmetric membership functions are used for the coefficients of the regression equation, has the best long-term performance in meeting variable demands.