Author(s): A. B. Celeste; K. Suzuki; A. Kadota
Linked Author(s): Koichi Suzuki
Keywords: Monthly reservoir operation; Optimization; Quadratic programming
Abstract: When the analysis of water resource systems involve a high number of decision variables and constraints, optimization models may help in providing operating alternatives which can be used by the water resource planners to assist their decision making. An optimization model based on mathematical programming was implemented for the determination of the optimal allocations of water from a multipurpose reservoir used for water supply and irrigation. The reservoir is part of the system that supplies the city of Matsuyama, in Japan. The objectives are to find the monthly releases of water that best meet the demands for city supply and irrigation and to maintain the reservoir storage as close as possible to a given target. A Quadratic Programming model was constructed and the gradient vector as well as the Hessian matrix of the objective function were determined. Several possible operations were carried out and the results were compared with fictitious simulations aiming all the demands to be met. The results show that the optimization model finds more reasonable operating policies than the simulations that try to meet all the demands without taking the future situation into account. Quadratic Programming also appears to be a good tool for solving the optimization model on a monthly basis.
Year: 2003