Title: 求償性隨機最佳化應用於整合性水資源管理之研究
Application of Stochastic Optimization with Recourse on Integrated Water Resource Management
Authors: 張良正
Keywords: 整合性水資源經理;求償性隨機最佳化;等定率模型;穩健模型;灰數模型;最好/最差模型;Integrated Water Resources Management;Stochastic Optimization with Recourse;Deterministic-equivalent Model;Robust Model;Grey-number Model;Best / worst-case formulation Model
Issue Date: 2011
Abstract: 台灣由於天然水資源有限,水源開發工程成本愈來愈高,為兼顧生態及環境保護原則,水資源管理應由傳統以水源開發為主轉為整合性水資源經理(Integrated Water Resources Management,IWRM ),其主要乃透過加強整合水資源需求面及供應面之管理,達到穩定供水的目的。整合性水資源經理之成本分析有別於傳統評枯方式,加入了利害關係人的參與以及方案間的互斥關係之考量,在考慮諸多方案間之競合及不確定性下,經常使用序率規劃之求償性隨機最佳化(Stochastic Optimization with Recourse)技術進行方案選擇、組合及排程,本計畫分別使用等定率模型(Deterministic-equivalent),穩健模型(Robust), 灰數模型(Grey-number)及最好/最差模型(Best / worst-case formulation)等4種模型來陳述求償性隨機最佳化求解方式。 本計畫主要目的為建立考慮不確定性與方案競合之整合性水資源經理方案決策模式,並以此模式分析桃園地區需求面與供給面之長、短期方案,優選最佳方案組合,期能以最小投資成本解決缺水問題。
Owing to the limited water resources, high developing cost and environmental concern, an Integrated Water Resources Management (IWRM) instead of conventional developing oriented approach should be applied on the future water resource management in Taiwan. IWRM considers both the supply development and demand management to provide sustainable water supply. The cost analysis of IWRM is different from conventional analysis and considers the stakeholder participation and strategies interaction. This study uses the stochastic optimization with recourse to optimize the best strategies combination under uncertainty. The study uses the four methodologies, deterministic-equivalent model, robust model, grey-number model and best/worse-case formulation to demonstrate the stochastic optimization with recourse. This study purpose is to develop an IWRM decision model to compute the best strategies combination which including long-term and short-term strategies. The model can obtain a optimal strategies combination with minimum expected cost and will be applied in Taoyuan area.
Gov't Doc #: NSC100-2221-E009-113-MY3
URI: http://hdl.handle.net/11536/99011
Appears in Collections:Research Plans