| 摘要: | 天然災害經常在日常生活中發生並造成生命與財產的重大損失，如地震、颱風、水災、火山爆發、土石流等。以地震災害為例，往往對災區造成相當大的破壞，不僅足以摧毀原有的交通與維生管線系統，且易對於災後之救援機具、救援車輛、救護車與運送物資車輛的運作造成嚴重的阻礙，進而影響救災效率與造成傷亡的增加。1999年的921集集大地震，不但暴露出政府危機處理能力的不足，更凸顯出大型災害緊急應變作業之重要性。以往傳統對於災後工程搶修與賑災物資配送，乃是由決策者依經驗進行工作隊與物資配送之指派作業，且將兩者分開並獨立處理。以目前實務的作法，忽略了兩者間之相關性，且缺乏系統性分析。其所得之配送計畫僅為一可行方案，難以確保為最佳及最有效率之方案。 緣此，本研究針對實務上兼顧災點搶修時間與物資配送時間最短之目標，並考量災後搶修與緊急物資配送之相關限制，構建一多目標之災後搶修工程與緊急物資配送模式，以供決策單位有效率決定災害搶修與緊急物資配送之順序，並於有限時間內找出最佳搶修效益和物資配送之計畫排程。本研究利用時空網路流動技巧，分別構建災點搶修時空網路與賑災物流時空網路。在網路的設計上，則以整數流動方式定式工作隊與物資於時空中的排程。二網路流動間再加上實務限制，以符合實際的搶修工程與配送物資作業。本模式定式為一多目標混合整數多重貨物網路流動問題，屬NP-hard問題。由於此模式的規模甚大，在短時間內難以求得最佳解，故本研究利用C程式語言並配合數學規劃軟體，發展一有效之啟發解法以有效求解問題。最後，本研究以類似921大地震規模之災害為例，測試模式及求解演算法的績效，進而提出結論與建議。 Natural disasters, such as earthquakes, typhoon, floods, volcanic eruption, mudflows and landslides, often occur in daily life and have significant devastating effects in terms of human injuries and property damages. For example, an earthquake will break down the traffic and lifeline systems and obstruct the operations of rescue machines, rescue cars, ambulances and relieves. As result, it could not only affect the rescue efficiency but also increase human injuries. The Chi-chi earthquake occurred in 1999 not only emerged the low efficiency of the government for dealing with the rescue operations but also revealed the importance of the emergency rehabilitation. In the past, the emergency rehabilitation and relief distribution were usually handled manually and separately by decision makers with their experiences, omitting the interrelation between emergency rehabilitation and relief distribution and lacking of systematic analyses. As a result, the solution could possibly be an inferior feasible solution. Therefore, we will develop a multi-objective programming model with the objective of minimizing the time length of the emergency rehabilitation and the relief distribution, subject to related operating constraints. The model is expected to help the decision maker to efficiently set the emergency rehabilitation and the relief distribution schedules under a limited time. We will employ network flow techniques including both emergency rehabilitation and relief distribution time-space networks in order to formulate the flows of working teams and relief commodities in the dimensions of time and space. Some side constraints will be set to these two networks according to the real constraints. The model is expectedly formulated as a multi-objective mixed integer multiple commodity network flow problem that is characterized as an NP-hard problem. Since the problem size is expected to be huge, we will develop a heuristic algorithm, which employs C computer language with the assistance of a mathematical programming solver, to solve the problem. To evaluate the model and the solution algorithm, we will perform a case study using data similar to that in Chi-chi earthquake. Finally, conclusions and suggestions will be given. |
