行為反應下航機誤點延滯擴散之研究

Abstract

摘 要 依照公開時刻表營運之航空公司，若時刻表中任一班次發生延誤，則下一班次將可能被迫發生延誤，造成誤點延滯之擴散。本研究應用機率與統計理論構建航機誤點延滯擴散模式，並加入行為反應之考量。以長期之觀點，行為反應乃起因於營運人員對於時刻表緩衝時間之預期，若在班表排程中加入緩衝時間，雖能提昇班表之準點績效，但營運人員因預期班次間有足夠緩衝時間(時間餘裕)，使得作業之效率降低，或造成不必要之延滯，亦將降低航空公司潛在之獲利機會。而以即時之觀點，行為反應乃由於實際營運中營運人員得知班機是否發生延滯而引起，若班機發生誤點之情形，營運人員將有可能加快作業之速度，盡量讓誤點之時間縮短。 本研究以隨機變數描述各班次航機之抵達時間、地面作業時間、與機門使用時間。依推擠性誤點可能發生之情況，分為三個部分探討：第一為同架航機、不同班次間之誤點延滯擴散，由於班次間使用同架航機，故下一班機是否準時起飛，直接受上一班機何時抵達機場，並完成地面作業時間之影響。第二為在機門之使用過程中，不同班機間之誤點延滯擴散，由於前後使用機門之飛機，不一定是同一架，故下一班機開始使用機門之時間，除了受到上一班機完成機門使用之時間影響外，亦受到本身航機抵達機門之時間影響，須同時考慮，本研究引入聯合機率分配之觀念處理此一問題。第三為轉機過程中，不同航機、不同班次間之誤點延滯擴散，而接駁班機何時能起飛，除受到欲轉機之旅客/行李登上接駁機之時間影響外，亦受到接駁機本身抵達機門時間之影響，亦可以聯合機率分配之觀念同時考慮。本研究構建之模式可推導出：下一班次離站準/誤點之機率、離站時間之機率密度函式、平均離站誤點時間、兩班次間平均之停等(浪費)時間等。 最後，本研究以C航空公司入境中正機場當天隨即飛往香港之穿越性班機資料作為樣本，檢定班機抵達誤點時間、地面作業延誤時間之機率密度分配，推算班機抵達與地面作業總誤點時間之機率密度分配，校估不同班表班距下之行為反應參數，並求得航機準時離站之機率、班機完成地面作業後至起飛前之平均等待(停等)時間、班機平均起飛誤點時間等重要指標，這些指標，將可作為航空公司分析班機誤點及擬定班表之參考依據。

ABSTRACT For airlines operating according to published time tables, whenever one flight is delayed, the following flights may be forced to be delayed accordingly, and consequently, results in flight-delay propagation. This study applies probability and statistical methods to model flight delay propagation effects, allowing for behavioral response. From a long-term view point, behavioral response is due to airlines’ estimate of the slack of the time table. Though putting more slack into the time table may improve flight punctuality, it reduces operating efficiency and entails unnecessary waits which cuts the profit potential of airlines. From the instant point of view, behavioral response is also brought upon by actual operators at that instance knowing the flight is delayed. In case of delayed flight, operators can increase their processing speed to minimize the delay. According to characteristics of flight delay propagation, we investigate the following three scenarios in the flight delay propagation model constructed in this research. The first is the flight delay propagation of the same airplane operating the consecutive flights. Because the two flights using the same airplane, the punctuality of the second flight is directly affected by the time the first flight arrives and the time it takes to unload passengers from the first flight and to load passengers for the second flight. The second is the flight delay propagation between different flights and different airplanes when they share the same gate. The punctuality of gate-use start time for the second flight directly depends on when the first flight finishes its gate use and when the airplane of the second flight arrives at the gate. The third scenario is flight delay propagation between different flights and different airplanes when transfers of connecting passengers and luggage are involved. This scenario is similar to the second case. At last, in this research, we use the data of the Airline C shuttle flights serving between Taipei and Hong Kong. In the case study, we first compute the probability distributions of time delays due to flight arrivals and time delays due to ground operations, and then derive the probability distribution of total time delay accounting for both factors. Second, we calibrate the value of the parameter representing the behavioral response due to flight schedules and gaps between flights. Finally, we calculate the probability of flights leaving gate on time, the average wait time of a flight from finishing ground operation to actual taking off, and average delay of the flight take-off time. The results of the study can be valuable references on both estimating flight delays and scheduling flights and slack time for airlines.