Determining an Optimal Action Portfolio for Water Resource Management by Using Stochastic Programming

Abstract

A regional water management system always contains long-term and short-term actions in practice. The uncertainty of short-term actions increases great complexity for evaluating the performance of action portfolio comprised with long-term and short-term actions. Developing an efficient methodology to define a cost-effective action portfolio is an important task. Hence, this study develops a novel decision model, the Stochastic Programming with Recourse Decision Model (SPRDM), to compute a cost-effective action portfolio. The effectiveness of SPRDM is verified by address a problem of water shortage and financial cost in Taoyuan City, Northern Taiwan. The results shown adding Kaotai Reservoir alone can fulfill the Taoyuan demand most of the time, and the remaining extreme water shortage events can be addressed by short-term and irrigation management actions additionally. When drought duration of water shortage event is longer than 200 days, using irrigation management action to address water deficit has lower expected costs than other short-term actions do. Although the study focuses on Northern Taiwan, the proposed model is applicable to other areas with an integrated water resources management framework.