Study on water sharing in a multi‐reservoir system using a dynamic programming‐neural network model

A dynamic programming–based neural network model is developed for analysing the water sharing between two reservoirs in a multi–reservoir system catering for irrigation. To study the water sharing between two downstream reservoirs from an upstream reservoir, a modified dynamic programming algorithm with three state variables and four decision variables is proposed. The operating policies are derived from the three state variable dynamic programming algorithm using a neural network. The new dynamic programming neural network model gives a very good performance for the multi–reservoir system case study considered. The performance of this model is compared with the improvised standard operating policy and constrained dynamic programming neural network model previously suggested.