求償性序率優選於水資源管理決策模型之研究

Abstract

整合性水資源管理（Integrated Water Resources Management, IWRM）係以永續發展為目標之概念，近代各國擴大引用於水資源管理，其強調因應缺水之方案組合與具成本效益之規劃決策，惟由於缺水之不確定性、方案成本與其增加供給或降低需求水量等特性，規劃決策是一個複雜的問題。本研究以求償性序率規劃為理論基礎，發展一個水資源管理決策模型（Stochastic Programming with Recourse Decision Model, SPRDM），可協助水資源管理者從諸多供給面和需求面之可能方案中，選擇具最小成本之長期與短期方案組合，以因應未來所有缺水事件。 本研究以假想案例進行SPRDM模型驗證與優選結果探討，並進一步以基隆地區水資源管理案例進行分析與探討，結果顯示： 1. 為因應未來所有缺水事件，相較於僅考量長期方案，長期與短期方案組合具有最小成本。 2. 長期與短期方案共同因應缺水事件，當遇更嚴峻的缺水事件，必需採取成本更高之短期方案。 3. 缺水事件與方案成本之不確定性將影響決策方案組合與合計成本。 4. 在台北自來水事業處未來仍持續供應基隆地區及該地區用水需求不超過預期的情況下，基隆地區水資源管理之3個情境優選結果顯示，節約用水計畫之成本在情境一設定為最低且為優選之長期方案，情境二與三優選之長期方案則為八堵抽水站與淨水場改善，考量節約用水計畫之成本不確定性，本研究建議八堵抽水站與淨水場改善及其相對的短期方案為決策方案組合，可因應所有缺水事件，該決策方案組合之成本在情境二與三分別約為每年新台幣2.37億元與3.02億元。

Integrated Water Resources Management (IWRM) is a concept rooted in sustainable development as a goal, and is expansively used for water management in many countries with emphasis on a set of actions in response to water shortage, as well as cost-effective planning decisions in recent years. However, with such factors as water shortage uncertainty, actions cost and water volume, making decision for a set of actions is a complex issue. This study proposes the Stochastic Programming with Recourse Decision Model (SPRDM) for the selection of long-term and short-term actions decision portfolio with the minimum cost. The actions cost include the combined cost of both long-term and short-term actions. The actions portfolio will handle all water shortage events and serve as a strategy reference for decision makers. This study undergoes SPRDM verification with hypothetical cases and explores the results, and its actual application to the water resources management in the Keelung region comes with findings as follows: 1. In comparison with merely considering the long-term actions, the decision portfolio of mixing long-term and short-term actions comes with minimum investment cost. 2. The set of long-term and short-term actions can handle water shortage events; in events of really severe water shortage, the short-term actions with much higher cost need to be adopted. 3. The uncertainty of water shortage events and action cost affects the decision portfolio and total cost. 4. Exploring the results of three scenarios for the water resources management in Keelung area, the Water Saving long-term action was selected in the optimal actions portfolio only for scenario I with least water saving cost while Badu Water Pumping Station and Water Purification Station Improvement action was selected for other scenarios. Therefore, considering the uncertainty of the Water Saving cost, this study suggests the Badu Water Pumping Station and Water Purification Station Improvement action with its associated short-term actions as the final decision portfolio. This portfolio can handle all shortage events and its total cost is approximately NT$237 million and NT$302 million for scenario II and III respectively.