- 學位論文
救援物資分配及運輸配送規劃之研究：以臺灣為例
Relief Allocation and Delivery Planning：A Study based on Taiwan
摘要
臺灣由於獨特的地理位置，因此易遭受天然災害的侵襲，造成島內居民生命財產的損失。若能有效對針對災害進行事前減災、備災，事後快速回應，將能降低天然災害對人民帶來的衝擊。而在救災時有良好的救援物資分配及運送系統，提供災民足夠的救援物資供給，將能支持災民快速從受災狀況中恢復。 本研究是針對災害管理中的應變階段，地方政府在面對管轄下的區公所提出救援物資需求後，如何分配自身有限救援物資到各區公所，並針對無法滿足的數量向中央政府請求物資的配送。此外地方政府在決定救援物資各區公所的物資分配量時，也需要同時考慮由地方政府統籌管理的運輸車運送路線及配送量。 因此本研究提出一個混整數規劃數學模型，來同時處理救援物資分配及運輸車使用和配送計畫，而本數學模型將追求最小化配送總成本、最小化總運輸時間，及救援物資分配的公平。且為了縮短求解時間爭取救災時效，本研究採用遺傳演算法，並以C#語言開發包含使用介面的救援物資分配派送系統，並進行災害情境測試，以分析本分配派送系統的適用性。
並列摘要
Due to the unique geographic location of Taiwan, natural disasters happen with high frequency and cause people’s live and wealth huge loss. A better plan in mitigation, preparedness and response for disasters will ease the impact and damage caused by nature on people. A well-planned relief allocation and distribution will supply enough relief to disaster-stricken people, and help them resume their normal life as soon as possible. In this study, we consider a relief goods allocation and distribution problem in which a local government is responsible for planning and managing the post-disaster rehabilitation for districts within its administration. After receiving relief goods demand from all of districts managed by the local government, the local government will plan how to allocate relief goods to meet demand. If relief supply is not insufficient to fulfill demand, the local government can send a request to the central government for asking supply for unmet demand. In addition to the allocation of relief goods, the local government also considers the routing schedule and delivery schedule of goods. Different types of vehicles with specific capability of transporting goods are considered. We formulate a mixed integer programming model to determine relief allocation and distribution. This model’s objective is to minimize delivery cost, transportation time and the maximum deviation of unmet relief goods among districts. A decision support system is constructed by using C#, and the system adopts a genetic algorithm to solve the problem in a large scale efficiently. Finally, we analyze the system’s applicability under different disasters scenarios.