Linear Decision Rules for Seasonal Hydropower Planning: Modelling Considerations

Abstract

The seasonal hydropower planning problem is presented as a robust optimization model using linear decision rules (LDR). Uncertainty is present in multiple inflow series and the price of electricity. The objective is to generate weekly policies that maximize profit. The LDR approximation is effective at reducing computational complexity, and is well-suited to multistage problems. By restricting the decision variables to be affine reactions of the realizations of the uncertain parameters, the original intractable problem is transformed into a tractable one with shorter computational time. This article is intended as an introduction to LDR for practitioners within the hydropower industry. The basic results will be presented, along with explanations of the ideas behind the method. An example case is given to illustrate the solution method.