Optimization Model for Allocating Water in a River Basin during a Drought

Abstract

Optimization of water use is a complex problem in a large scale river basin. One of the most important approaches in optimizing water use in a river basin is to find the relationship between water demand and water supply. The parameters that affect demand, supply, and the methods of evaluation of such elements are discussed in this study. Also, a method is presented for providing objective and constraint functions from considering these effects. Fuzzy logic theory is used to modify the stochastic dynamic programming (SDP) method such that an optimization model is developed for allocating water and can be defined as the “stochastic fuzzy dynamic programming (SFDP)” method. This method is applied to optimize water use in the Kor and Seevand river basins, located in the Bakhtegan watershed, Fars, Iran. The primary water resources management consisted of the variability ranges of decision variables such as release from Doroodzan Dam and reservoir storage and was also used for allocating water in these river basins based on the SDP method. Therefore, in the present study, these variability ranges are obtained based on historical data, and divided into several record classes. Optimum class of release, a case of the record classes, was obtained from the optimization model for each month during the past 4 out of 25 years. Although, the SFDP method can be used in optimizing water allocation during each period, the method is structured and discussed only during the drought periods (4 years). Later, a comparison was made between optimum classes and record classes that were operated during the primary water resources management. During this period, the SFDP method reduced the difference between the release from the dam and the total water demand of the river basin. Therefore, approximately a 27% improvement in adaptation between release and demand could be attained. Finally, if the decision maker makes the decision for the release from the dam that is optimal according to our objective function, the reliability of reservoir operating can be increased by 51% during future droughts.