考量航空公司機門偏好之機門指派最佳化模式

Abstract

隨著全球機場逐漸朝向民營化及公司化之趨勢，政府的角色已由管理者演變為監督者。而有效率且公平的機門指派是機場營運的重要工作之一，且基於政府觀點之機門指派核心邏輯已逐漸轉變為融入航空公司的需求及觀點。航空公司於機場載運旅客，可能因其所需設施、合作之地勤業者所在位置或基於油耗、減少到離航班之拖機次數等相關場站費用，而有所偏好特定之機門。 爰此，本研究構建一機門指派模式，同時以旅客總步行距離最小化及航空公司對機門偏好最大化為目標，反應航空公司實務營運，此雙目標數學規劃模式使用分支界定法與LINGO程式進行求解。另為了探討模式之適用性，本研究以臺灣桃園國際機場為例，並分就淡旺季及尖離峰時段之實際旅運需求資料進行測試，進而分析比較在不同實際需求大小的求解結果。 結果顯示本模式在各種不同情境需求下，皆能求解妥適且合理之解決方案。此外，當航空公司對機門偏好重疊性愈高，愈會有些航班在尖峰時段將被指派至外機門及非偏好機門，且偏好機門的利用率也會低於非偏好機門，亦屬合理，顯示本模式具備可應用性。

With the trend of the privatization or corporatization of airports worldwide, the role of governments is changed to be a supervisor instead of an operator. Efficient and fair gate assignment is one of key task of airport operation. The core logic of gate assignment originally mainly based on governments’ viewpoints has been gradually changed to value the views and requirements of airlines. Airlines may prefer specific gates because of the location of required facilities, cooperated ground handling service team and fuel consumption conservation, reduction of the number of flight transferring and other related expenses. Based on this, this study develops a gate assignment model to simultaneously minimize the total walking distance of passengers and maximize airline gate preferences. The proposed bi-objective mathematical programming model is then solved by Brand-and-Bound method and Lingo software. To investigate the applicability of the proposed model, a case study on Taiwan Taoyuan International Airport is conducted. The traffic demand of peak and off-peak seasons and hours are applied to compare the gate assignment results of various sizes of real demand. The results show the proposed model can solve the comprised and reasonable solutions under various demand scenarios. Additionally, when the preferred gates of airlines are highly overlapped under peak hours, some aircrafts will be assigned to remote gates and non-preferred gates. The utilization ratios of preferred gates are also lower than those of nonpreferred gates as expected, suggesting the reasonability of the assignment results and the applicability of the proposed model.