Optimization and adjustment policy of two-echelon reservoir inventory management with forecast updates

A finite-horizon, periodic-review inventory model with inflow forecasting updates following the Martingale Model of Forecast Evolution (MMFE) in multiresevoirs is considered. This model introduces a new method of determining an operating policy in which the policy is based on the Dynamic Programming (DP) model with a physical equation and a recursive equation. Some important constraints are described for the multireservoirs operation. We propose two methods to adjust the decision variable during the optimization. The adjustment policy is essential to improve the profit and decrease the shortage in multireservoirs management. Finally, to assess the effectiveness of the policies, the model is compared with other models and is applied to the Chinese South–North Water Diversion project.

► Finite-horizon, periodic-review inventory model with forecasting update is considered.
► It introduces a new method of the dynamic programming model in multireservoirs.
► We propose two methods to adjust the decision variable during the optimization.
► The model is compared with other models and is applied to SNWD project.