- 學位論文
災害管理之應急物流合作與資源定位配置
Emergency Logistics Collaboration and Resources Location-allocation for Disaster Management
摘要
雖然供應鏈合作已被證實其在幫助成員達成共同目標是基本的，但目前應急物流文獻所使用之模型上仍未將供應鏈合作的觀念導入，尤其是供應鏈合作的應用與優點在應急物流管理上未出現在文獻中，以上為本文探討此類研究議題的動機。本研究企圖系統地應用合作觀念探討三個有關應急物流的議題。三個研究議題為: 救災伙伴選取與資源配送(relief supplier selection and distribution, RSSD)；緊急供應網路設計(emergency supply network design, ESND)；自組緊急應變 (self-organized emergency response, SOER)。針對三個議題所提出之方法如下所示： (1)本研究針對RSSD提出一個救災供應合作方法。此方式包括兩層函數: (a) 二階段聚類方法辨示潛在救災伙伴; (b)使用隨機規劃模型決定多類型的資源供給量，以最小化供需不平衡效應在災害應變時期的影響。這方法特點是在合作下找到潛在救災伙伴與最小化供需不平衡效應。模型測試結果指出供需不平衡效應在救災伙伴聚類下相較於未下聚類有明顯優勢。 (2)研究對ESND提供一個集中型緊急供應網路的設計方法，此方法採用三階段的混合目標整數線性模型，其整合了避難所、醫療、及配送等三個子網路。方法的特色: (a)採用需求導向，子網站設計的順序是避難所、醫療、及配送，其中在後者之子網路設計基於前者之子網站。(b) 模型不僅包含傳統目標式，如供給端重視的最小化總移動距離及作業成本，也包含需求端重視的最小化心理成本。模型測試結果指出使用本方法所設計之集中型緊急供應網路優於分散型網路，尤其是在設計配送子網路上。 (3)在近期的文獻中，有關決策單位收集災區外部物流資源執行緊急應變的議題已廣泛被討論，然而，著眼於災區內部自我組織與動員物流資源之議題是缺乏的。研究針對SOER提出一個自我組織緊急物流資源動員系統(SELRMS)，此系統動員及整合災區內的物流資源以因應在緊急應變時期的需求。系統形成主要使用混合目標線性規劃模型，此模型不僅考慮傳統目標式，如最大化生存人數及最小化缺少供給量，也考慮需求導向之目標，如最小化屍體曝露風險以及災民的心理成本。數值分析研究一個真實的災難，以驗證系統的可應用性及優點。分析結果指出具可再配置人員功能的系統相較於基本型系統(無可再配置人員功能)，較能有效率地運用災區的救災能量。 本研究期望藉所提出減輕災害影響之方法對相關研究有所貢獻。
並列摘要
No programming model developed by study has been found to incorporate the concept of supply chain collaboration into emergency logistics although supply chain collaboration has been proved to be is essential in assisting in the achievement of joint goals of supply chain members. Particularly, the potential application and advantages of collaboration in the emergency logistics management have never been discussed and investigated in the previous literature. This is also the major motivation to address such a dissertation in which this study intends to comprehensively and systematically discuss the three research issues of emergency logistics applying the concept of collaboration. Three research issues are relief supplier selection and distribution (RSSD), emergency supply network design (ESND), and self-organized emergency response (SOER). The proposed methods for these three research issues are as follows: (1) This study proposes a relief supply collaboration approach for RSSD. This proposed approach involves two functions: (a) a two-stage relief supplier clustering mechanism for time-varying multi-source relief supplier selection, and (b) the use of stochastic programming model to determine a multi-source relief supply that minimizes the impact of relief supply–demand imbalance during response phase. The distinctive features of this proposed approach are to identify the potential relief suppliers and to minimize the imbalanced supply–demand impact under relief supply collaboration. Model tests are conducted to demonstrate that relief supply collaboration with grouped relief suppliers has a significant benefit of alleviating the impact of imbalanced relief supply–demand, relative to collaboration with ungrouped ones. (2) This study proposes a method for designing a seamless centralized emergency supply network by integrating three sub-networks (shelter network, medical network, and distribution network) for ESND in the aftermath of a disaster. The proposed method primarily involves three stage multi-objective (travel distance minimization, operational cost minimization, and psychological cost minimization), mixed-integer linear programming models. The three sub-networks are designed using the proposed programming models. The distinctive features of the proposed method are as follows: (a) the proposed method is demand-driven. The order of the designed sub-networks is shelter, medical, and distribution, with the connections of the latter networks based on the arrangements for the former; (b) the objective functions of three stage programming models include not only traditional objectives (i.e., travel distance minimization and operational cost minimization), which supply-side members focus on, but also minimizing the psychological cost experienced by demand-side members. Model tests are conducted to demonstrate that the superiority of a centralized emergency supply network designed by the proposed method over a decentralized one, especially with regard to distribution network design. (3) Emergency response out sourcing external logistics resources dominated by decision making units outside affected areas (e.g., central government and international NGOs with logistics divisions) has been investigated widely in recent literature. However, research that focuses on self-organization and mobilization of internal logistics resources inside affected areas is scarce. This study proposes a self-organized emergency logistics resource mobilization system (SELRMS) which permits mobilizing and integrating logistics resources inside an affected area for SOER during the emergency response period. The resource-integration problem of the proposed SELRMS is formulated using a multi-objective linear programming model, which takes into account not only traditional objectives (i.e., survival maximization and undersupply cost minimization), but also demand-oriented objectives (i.e., corpse exposure risk minimization and psychological cost minimization). To demonstrate the applicability and advantages of the proposed system, numerical analysis aiming at a real disaster case is conducted. Analytical result demonstrates that the extension system (i.e., thee system with relief worker reallocation function) facilitates efficient assignment of relief workers and improves objective values rather than basic one (i.e., the system without relief worker reallocation function). Author hopes that this dissertation contributes to related research by developing the methods applicable for alleviating disaster impact in affected areas.