隨機需求下機場護送任務指派之研究

AIRPORT ESCORT ASSIGNMENT WITH STOCHASTIC DEMANDS

近年來，隨著機場護送任務的數量逐漸增加，如何適當指派護送任務以服務旅客的需求已成為機場營運重要的課題之一。目前護送任務指派方式仍大都仰賴決策者經驗，缺乏系統最佳化規劃，常導致不佳的指派，造成有時延遲或取消護送任務，有時卻又浪費人力等情況。因此本研究考量現實中護送任務需求的隨機性及相關的營運限制，以總護送員護送作業成本期望值最小為目標，構建一隨機需求下護送任務指派模式，期能幫助國內機場管理單位或相關業者有效地進行實務的機場護送任務指派。本研究應用網路流動技巧，構建一護送員作業網路，並以數學規劃方式建立模式，其可定式為含額外限制整數網路流動問題。本研究並配合滾動分析，在營運過程中動態地指派護送任務。為初步測試模式的應用績效，本研究參考國內某國際機場之相關護送任務資料為例，進行範例測試，且針對數個重要參數進行敏感度分析，結果良好，顯示本研究模式應可以為機場管理單位或相關業者之實務參考。

In recent years, the number of airport escorts have gradually increased. How to appropriately assign escorts to traveler requests has become one of the important issues in airport operations. However, the current method of escort assignment is mostly relied on decision makers’experiences, without a perspective of systematic optimization, usually resulting in inferior assignments with delayed/canceled escorts occasionally, while having waste of manpower sometimes. Therefore, we consider the stochasticity of escort demands and related operating constraints to develop an escort assignment model with stochastic demands, with the objective of minimizing the total expected escort operating cost. It is expected to be useful references for airport authorities or related carriers to effectively perform airport escorts in practice. We employed network flow techniques to construct an escort network and adopt mathematical programming method to establish the model which can be formulated as an integer network flow problem with side constraints. We also adopt a rolling analysis to dynamically assign escorts in operations. To preliminarily evaluate the model, we perform a case study using data from an international airport in Taiwan with a number of sensitivity/scenario analyses. The test results are good, indicating that the model could be useful for airport authorities/related carriers for their practices.