Premium
The implicit stochastic model for reservoir yield optimization
Author(s) -
Simonovic Slobodan
Publication year - 1987
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/wr023i012p02159
Subject(s) - computation , yield (engineering) , mathematical optimization , resource (disambiguation) , multivariable calculus , reliability (semiconductor) , function (biology) , reservoir computing , water supply , computer science , mathematics , engineering , algorithm , artificial neural network , materials science , metallurgy , computer network , power (physics) , physics , quantum mechanics , control engineering , evolutionary biology , machine learning , environmental engineering , recurrent neural network , biology
The implicit stochastic model is aimed at solving the specific problem of the optimal reservoir yield when the demand is not known. The model is created to assist in the long‐term comprehensive water management planning. The presented treatment of the reservoir as a water resource for satisfying demand differs from the present practice. Knowing the character of different water demands, represented by relative demand coefficients, the implicit stochastic model provides planners with the optimal value of reservoir yield. The lack of strong and reliable economic criteria, which is common for development countries, is in this approach replaced by optimizing the reliability of satisfying water demand from the reservoir. A three‐level algorithm is proposed for the reservoir yield computation. At the first level, the simulation is used for computing the value of objective function. At the second level, computation of the seasonal reservoir operating rules is presented. The approach used for deriving the reservoir operating rules is based on the nonlinear unconstrained multivariable search of M.J.D. Powell. The third level is used for estimating the single multipurpose reservoir yield. The Fibonacci search procedure is used for the optimization of the reservoir yield at this level. The model is used in reservoir analysis for the water resources master plan of the Republic of Serbia, Yugoslavia. In this case, 49 reservoirs Were analyzed.