Hindawi

Journal of Advanced Transportation

Volume 2019, Article ID 7069380, 8 pages
https://doi.org/10.1155/2019/7069380

Research Article
A Method of Reducing Flight Delay by Exploring Internal
Mechanism of Flight Delays

Yakun Cao ,1 Chenping Zhu,1 Yanjun Wang ,2 and Qingyun Li2

1
College of Science, Nanjing University of Aeronautics and Astronautics, Nanjing 211100, China
2
Civil Aviation Institute, Nanjing University of Aeronautics and Astronautics, Nanjing 211100, China

Correspondence should be addressed to Yakun Cao; [email protected]

Received 19 March 2019; Revised 30 July 2019; Accepted 30 August 2019; Published 30 December 2019

Academic Editor: Luigi Dell’Olio

Copyright © 2019 Yakun Cao et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper explores the internal mechanism of flight departure delay for the Delta Air Lines (IATA-Code: DL) from the viewpoint
of statistical law. We roughly divide all of delay factors into two sorts: propagation factor (PF), and nonpropagation factors (NPF).
From the statistical results, we find that the distribution of the flight departure delay caused by only NPF exhibits obvious power
law (PL) feature, which can be explained by queuing model, while the original distribution of flight departure delay follows the shift
power law (SPL). The mechanism of SPL distribution of flight departure delay is considered as the results of the aircraft queue for
take-off due to the airports congestion and the propagation delay caused by late-arriving aircraft. Based on the above mechanism, we
develop a specific measure for formulating flight planning from the perspective of mathematical statistics, which is easy to implement
and reduces flight delays without increasing operational costs. We analyze the punctuality performance for 10 of the busiest and
the highest delay ratio airports from 155 airports where DL took off and landed in the second half of 2017. Then, the scheduled
turnaround time for all flights and the average scheduled turnaround time for all aircraft operated by DL has been counted. At last,
the effectiveness and practicability of our method is verified by the flights operation data of the first half of 2018.

1. Introduction preferential attachment [16, 17]. On the other hand, in Ref.

[18], the author proposed a SPL’s model with a parameter
Flight delay is one of the major issues in aviation systems all which controls the relative weights between the power-law
over the world. Such delay events downgrade the functioning and exponential behaviors. Empirical investigation for many
of airlines and cause tremendous loss in human life, economy real world networks [19–21] also shows SPL distribution.
and traffics [1, 2]. To alleviate the harm of flight delay, consid- These work provide an effective theoretical support for us to
erable work has been done [3–9]. Actually, the air transporta- explore the internal mechanism of flight delay and propose
tion system is a rather complex system, which have been effective measures to alleviate the harm of flight delay.
traditionally described as graphs with vertices representing There are many factors that cause flight delay, the Bureau
airports and edges direct flights during a fixed time period of Transportation Statistics (BTS) classifies them into five cat-
[10]. These graphs are called aviation network. Recently, many egories [22]: (1) aircraft arriving late, (2) national aviation
research has been carried out from the viewpoint of complex system (NAS) delay, (3) air carrier delay as a result of crew,
network [11–13], which propose almost all kinds of aviation baggage loading or maintenance problems, (4) extreme
network features. weather conditions such as hurricanes or blizzards and (5)
Many networks in nature display rather complex struc- security-related delays. If one flight is delayed, then a subse-
tures, that often seem random and unpredictable. Barabási quent flight might also be delayed because it is awaiting that
and Albert discovered that many realistic networks [14, 15] inbound aircraft. This kind of delay is called propagation delay
exhibit the scale-free feature, which the vertex connectivity [3, 5, 22–25], which is also quite substantial (more than one-
follows a PL distribution. The fundamental mechanisms lead- third of the delays) [3]. On the other hand, since the schedule
ing to the PL distribution are considered to be growth and of one aircraft is quite tight, the on-route absorption of
2 Journal of Advanced Transportation

Table 1: The information for DL airline in the second half of 2017.

Airline Number of flights Number of airports Number of aircraft Flight departure delay ratio (delay more than 15 minutes)
DL 470267 155 841 11.97%

departure delay of the last flight is very limited and the delay 0.01
in subsequent flights is relatively predictable, while the delay
caused by NPFs is hard to predict. Thus, quantitative research
1E-3
of propagation delay is great significance, which helps to come
up with solutions.
In order to alleviate the delay of propagation, researchers 1E-4
have proposed to modify schedule departure time so as to

p (l)
re-allocate the existing slack in the flight schedule [3, 6, 26–
28]. These studies share a similar research methods: they allow 1E-5
c ≈ 132.43
schedule departure time to vary within a time window, then β ≈ 25.83
establish an objective function with several constraints, and γ ≈ 2.74
1E-6 2
R ≈ 0.999
finally obtain the optimal solution. They focus on the impact
of schedule modification on system performance to maximize
the utilization of aviation resources. But we are more con- 1E-7
10 100 1000
cerned about how to reduce flight delay ratio and hope to
Delay l (min)
propose the concrete practicing method. In the follow, we
–γ
propose a specific implementation method, not an objective DL p(l) = c(l+β)
function, although we used the same idea as the previous stud-
ies, that is, modify schedule departure time. We take advantage Figure 1: Log–log plots of PDF of the flight departure delay.
of the predictability of propagation delay and assume that
there is no newly formed delay (delay caused by NPFs) after
changing the plan, the effectiveness and practicability of our delay distribution shows attenuation trend, which is faster than
method is verified by the flights operation data of the first half the linear attenuation in double logarithmic chart. Therefore,
of 2018. we consider the departure delay distribution is well approxi-
The structure of this paper is organized as follows: Section mated by SPL:
−𝛾
2 presents a statistical law for airline of DL and explores inter- 𝑝(𝑙) = 𝑐 ⋅ (𝑙 + 𝛽) . (1)
nal mechanism of flight delay. Section 3 contains analysis for
operation performance evaluation of different airports and Shown in Figure 1, the fitting function 𝑝(𝑙) of SPL can
statistical results of the scheduled turnaround time for all describe the empirical data very well. Statistical data shown
flights and the average scheduled turnaround time for every as black filled circles, while red fitting line in panel describes
aircraft. And the specific method is put forward. Section 4 the fitting result of Formula (1), in which the corresponding
presents and discusses the empirical results. In Section 5, con- parameters 𝑐 ≈ 132.43, 𝛽 ≈ 25.83, and 𝛾 ≈ 2.74. The constants
clusions and some hints for future research are given. of 𝑐, 𝛽, and 𝛾 are estimated in the way of the least square fitting,
and the goodness of fit is about 𝑅2 ≈ 0.999.
To explore the internal mechanism of flight departure
2. Statistical Law and Internal Mechanism delay, we first investigate the factors causing flight delays. As
shown before, delay factors include five categories, we consider
We collect primary records of flights operation from July 1, these five kinds of factors can be roughly divided into two
2017 to December 31, 2017 for the Delta Air Lines. The data sorts: the propagation factor (PF), i.e., category (1) aircraft
of flights operation were downloaded from the website of the arriving late, and the nonpropagation factor (NPF) which
Bureau of Transportation Statistics (BTS) [29]. Our analysis include all other four. Flight delays caused by NPFs are more
focuses on the departure delay rather than the arrival delay, accidental, while delay propagation has more direct relevance.
because the arrival delay is approximately linearly related to Delay propagation occurs when late arrivals at an airport cause
the departure delay [30]. In general, the departure delay is late departures, which in turn cause late arrivals at the desti-
commonly measured as the difference between the scheduled nation airports. In general, the air traffic controller will set
and the actual flight departure time. The Federal Aviation appropriate turnaround buffer time to prevent propagation
Administration (FAA) defines the flight departure delay as the delay when formulating flight planning [7], although this
flights departure at least 15 minutes behind schedule. The method reduces revenue-marking flight time and incurs
detail information for primary data is listed in Table 1. schedule time costs. From the follow statistical results, we find
In order to vividly describe the flight delay, we plot in that current measure of setting buffer time does not play a
Figure 1 the probability distribution function (PDF) of the prominent role.
departure delay and set 𝛥𝑙 (statistical interval of PDF) of PDF Actually, a key challenge to explore the internal mecha-
is equal to 15 minutes. Clearly, we notice that the departure nism of flight delay is extracting effective information from
Journal of Advanced Transportation 3

0.01
0.01

1E-3
1E-3

1E-4
1E-4
p (l)
2

1E-5

p (l)
2
1E-5
c ≈ 1.29
2
1E-6
γ ≈ 2.02
2
1E-6 2
R ≈ 0.998
1E-7
1E-7
10 100 1000
Delay l (min) 10 100 1000
Ttur = 30 mins Ttur = 40 mins
Delay l (min) –γ
Ttur = 50 mins Ttur = 30 mins p (l) = c .l 2
2 2

(a) (b)

0.01
0.01

1E-3
1E-3

1E-4
1E-4
p (l)
2

p (l)

1E-5 c ≈ 0.84
2

2 1E-5
γ ≈ 1.96 c ≈ 0.60
2 2
1E-6 2
R ≈ 0.997 γ ≈ 1.89
1E-6 2
2
1E-7 R ≈ 0.996

1E-7
10 100 1000
Delay l (min) –γ2
10 100 1000
Ttur = 40 mins p (l) = c .l Delay l (min)
2 2 –γ2
Ttur = 50 mins p (l) = c .l
2 2

(c) (d)

Figure 2: (a) Log–log plots of the departure delay (remove the delayed flights causing by PF) distribution 𝑝2 (𝑙) with 𝑇𝑡𝑢𝑟 = 30, 40, and 50
respectively. (b)–(d) are the curve fitting of the departure delay distribution with 𝑇𝑡𝑢𝑟 = 30, 40, and 50 respectively, the parameters value and
the goodness of fit is shown in plots.

the raw data. Because the existing data do not provide direct speaking, the larger the passenger capacity of the aircraft, the
information to distinguish between the different types of delay longer the necessary ground service time). That means if the
factors [23]. The other reason is that flight delay may be not time between the last actual arrival and the current schedule
merely attributed by a late arrival of the flight immediately departure is less than 30–50 minutes, it can be attributed to
preceding it, but also be attributed by one or more other factors propagation delay.
(NPFs). In order to quantitatively study the propagation delay To explore the impact of PF on the statistical law of the
and simplify the cause-explanation of late-arrival in the pres- departure delay, we remove the departure delay causing by PF
ent work, we consider that: a delayed flight with the time from the raw data. Since the data we collected without the
between the last actual arrival and the current schedule depar- information about the passenger capacity for different aircraft,
ture less than 𝑇𝑡𝑢𝑟 is attributed by PF. We know that the sched- we plot the departure delay (remove the delayed flights causing
ule turnaround time is consisting of two portions, namely the by PF) distribution 𝑝2 (𝑙) by setting 𝑇𝑡𝑢𝑟 = 30, 40, and 50 for all
schedule buffer time and the standard ground service time aircraft in Figure 2.
[31]. For different types of aircraft, the required standard It exhibits a PL distribution instead of a SPL distribution,
ground service time is about 30–50 minutes (generally given by
4 Journal of Advanced Transportation

Table 2: Statistical results of 10 airports with the highest delay ratio.

Number of delayed flights (delay more than

Airports (IATA-code) Number of flights Total delay (in minutes) Delay ratio (%)
15 minutes)
SFO 6753 1298 82510 19.22%
EWR 2779 514 39920 18.49%
JFK 14472 2536 197146 17.52%
LGA 11289 1795 144010 15.90%
MIA 4668 727 54702 15.57%
PBI 2936 455 29964 15.49%
ORD 4200 644 44935 15.33%
BOS 7827 1170 97251 14.94%
LAX 16661 2464 139844 14.7%
SEA 12102 1679 94217 13.87%

number of delayed flights caused by PF, and the fewer the

number of delayed flights caused by NPFs. On the other hand,
0.01 the smaller the 𝑇𝑡𝑢𝑟 value, the better the fit of the curve using
the PL function.
1E-3
In the statistical process, we use different thresholds 𝑇𝑡𝑢𝑟
to obtain PL distributions, which shows that the distribution
of flight delay caused by NPFs does exhibit the characteristics
p (l)
3

1E-4 of PL distribution. To understand the origin of this observed

PL distribution, we have to realize that the airport runway
restrictions and the take-off queue size as the significant causal
1E-5 factors that affect the actual departure time [32]. One delayed
flight caused by NPFs, such as extreme weather, the flights
behind this at the same airport usually delay too. When emer-
1E-6
10 100 1000 gencies return to normal, the waiting aircrafts' takeoff is a
Delay l (min) queuing process. Therefore, the distribution characteristic
SFO PBI shown in Figure 2 can be regarded as the consequence of a
EWR ORD decision-based queuing process [17, 33, 34]: when some per-
JFK BOS ceived priority has been executed, the time of the planes wait-
LGA LAX ing for take-off will show the characteristic of PL, with most
MIA SEA
flights rapidly take off, whereas a few experience very long
Figure 3: Log–log plots of departure delay distribution 𝑝3 (𝑙) for the waiting times.
10 busy airports with the highest delay ratio. Therefore, the mechanism of SPL distribution of flight
departure delay is considered as the results of aircraft queue
for take-off due to the airports congestion and the propagation
delay caused by late-arriving aircraft.
𝑝2 (𝑙) = 𝑐2 ⋅ 𝑙−𝛾2 , (2)
where 𝑐2 is a constant and 𝛾2 is a constant parameter of the 3. Method
distribution known as the exponent or scaling parameter. We
obtain the value of 𝑐2 and 𝛾2 by the way of the least square According to the previous mechanism of the flight delay, we
fitting (after taking the log of the two sides and 𝛾2 becomes the can deal with the flight delay from two aspects, namely the
slope of the line). As shown in Figure 2, the main part of the airports congestion and the propagation delay. The most effec-
distributions fit well with the fit function of Formula (2), while tive way to reduce queuing time is building multiple airport
the tail of distributions (larger delay) do not appear to be cap- runways. However, it is a huge investment. From the perspec-
tured by it. However, the goodness of fit of R2 for all distribu- tive of statistics, a new method is developed to improve the
tions with different 𝑇𝑡𝑢𝑟 is bigger than 0.99. From the data, we flight on-time performance. This method consists of two
find that the number of delayed flights with delay l bigger than stages: (1) data statistics and summarization; (2) implemen-
500 minutes is about 400–500, accounting for only 0.085– tation steps.
0.106% of the total number of flights. The fact that the scaling
spans close to two orders of magnitude, from minutes to hours, 3.1. Data Statistics and Summarization. Due to the airports
indicates that most flight delays (70.51% for DL) are within congestion, delay originating from these airports spreads to
less than one hour. With the increasing of the 𝑇𝑡𝑢𝑟 value, the downstream flights. So the operation performance of airports
value of the distribution function is smaller. Obviously, the plays a vital role in the punctuality ratio of airlines. The data
longer the necessary ground service time, the more the that we collected not only contains the message of time for
Journal of Advanced Transportation 5

250

80000
200

N (number of aircraft)
N (number of flight)

60000
150

40000
tur = 75.3 100
σ = 92.9
20000 50

0 0
0 40 80 120 160 200 240 280 320 360 400 0 20 40 60 80 100 120 140 160 180 200
The scheduled turnaround time (min) The average scheduled turnaround time (min)
(a) (b)

Figure 4: (a) Number distribution of the schedule turnaround time for 347073 flights. (b) Number distribution of the average schedule
turnaround time for a total of 728 aircraft.

departure and arrival, but also the carrier, tail number of and where required, catering and cabin cleaning procedures.
aircraft and the airports for departure and arrival. Next, we This measure is associated with airport operational efficiency
assess the operation performance of each airport and compute and is used to improve the planning of flight connectivity and
the scheduled turnaround time for all flights and the average the robustness of flight plan. In our method, we will modify
schedule turnaround time for every aircraft. the existing flight schedule and redistribute part of the sched-
While recent studies on air traffic delays focus primarily ule buffer time in the flight schedule without changing total
on operation performance for the different airlines [22, 35], slack time of the day and total daily number of flights.
we are interested in operation performance for the different In order to properly reset the slack, we count the scheduled
airports. As we know, airports are distributed in different loca- turnaround time for all flight and the average scheduled turn-
tions, the punctuality ratio for different airports are very dif- around time for all aircraft operated by DL in the second half
ferent due to the weather conditions and other regional of 2017. Since there are typically no flights between 0 and 6
factors. From our statistical results, we find there are 44 air- o’clock, we do not take into account this longer time when
ports which have more than 2,000 taking-off flights in the calculating the scheduled turnaround time. On the other hand,
second half of 2017 and 10 of 44 airports with the highest delay records available in BTS are not always complete for all aircraft.
ratios are reported in Table 2. We can see that, airport of SEA To promote the quality of statistics, we take 100 flights within
has more delayed flights than BOS, but the total delay is 6 months as the filtering threshold, which means that aircraft
smaller. That means the flight delay of airport BOS is mostly with their taking-off records smaller than 100 will not be
larger than SEA, so delay at airport BOS will have a greater counted into our statistics in the present work. After filtering,
impact on subsequent flights. a total of 728 aircrafts are counted, and the total number of
Initial delays affect the downstream flights, but small turned around for these aircraft is 347073.
delays do not have much impact due to the scheduled turna- The scheduling of aircraft turnarounds is a consequence
round buffer time. The study of delay distribution for various of both the operational policies and the scheduling strategies
airports is necessary, not only delay ratio. In Figure 3, we com- of an airline. For different airlines, the average scheduled turn-
pare the flight departure delay distributions of 10 airports. around time is quite different, Southwest Airlines in the USA
From Table 2, we know that airports of JFK, LGA, LAX, and shows a low average aircraft turn time of 17 minutes and
SEA concentrate a large part of Delta Airline’s flights, but the United Airlines an average turn time of 50 minutes [36]. In
characteristics of their delay distributions are not very different Ref. [36], we know that Delta Airlines shows an average turn-
from each others. The shape of the delay distribution of differ- around time of 46.7 minutes, in which the database includes
ent airports is similar, but small difference can only be information from September 1987 to May 1994. According to
observed when one focuses on EWR airport. The EWR airport our statistics, the average scheduled turnaround time 𝑡𝑡𝑢𝑟 of
shows a bias toward larger delays and may have a greater all flightsis about 75.3 minutes and standard deviation 𝜎 is
impact on subsequent flights than other airports. about 92.9. This shows that the scheduled turnaround time of
The insufficient schedule turnaround time is another flights has increased greatly nowadays, it is particularly advan-
important factor for causing the propagation delay. The sched- tageous to our method of redistributing part of the schedule
ule turnaround time stands for the time spent by an aircraft buffer time. Number distribution of the schedule turnaround
on ground from scheduled arrival to scheduled departure from time is shown in Figure 4(a), almost all flight’s scheduled turn-
the gate, which is used for an aircraft to absorb last flight delay, around time is longer than 30 minutes. So we set the minimum
complete full off-loading and loading maintenance of aircraft necessary turnaround time to be 30 minutes in our method.
6 Journal of Advanced Transportation

Time

Space

t1 tact1
Airport 1

Δt Flight 1
t1' t2 tact2
Airport 2
tact1'
tbuf2 tser2
Flight 2
ttur2
t2' tact2' t3
Airport 3
tbuf3 Ttur
Flight 3
ttur3
t3'
Airport 4

Figure 5: Time–space graph of one aircraft’s route.

In Figure 4(b), we can see that almost all aircraft’s average 30000
scheduled turnaround time is about 50–140 minutes. If we set
the necessary turnaround time too large, then the change to 25000
the flight plan is small, and the effect of restraining delay prop-
Number of delayed flights

agation will not be obvious.

20000

3.2. Implementation Steps. The overall approach is based

15000
on the flight delay mechanism where newly formed delays
usually occur at busy airports due to airport/airspace capacity
constraints and they spread to downstream flights by the same 10000
aircraft. From our data, it is possible to trace the propagation
of delay from airport to airport: if a particular aircraft is 5000
scheduled to fly from airport A to airport B and then to airport
C and departs from A with a long delay, part or all of that delay 0
0 60 120 180 240 300 360 420 480 540
will be propagated downstream and result in departure delay Delay l (min)
at B and, possibly, subsequently at C. In this section, we will
develop a new method for formulating flight planning by using Ttur = 30 mins Ttur = 50 mins
the previous statistical results. Ttur = 40 mins The first half of 2018
Since the newly formed delay was hard to predict when
we formulated the flight planning, we simply assume that Figure 6: Comparison of the number of delayed flights before and
flights departure from these 10 of the highest delay ratio air- after adjusting flight planning.
ports mentioned above will experience this kind of delay.
Actually, we cannot reduce the newly formed delay by opti-
mizing flight plans, but we can mitigate the propagation effects Specific measures are as Figure 5, where 𝑡𝑖 means schedule
of last flight delay by postponing the scheduled departure time departure time, 𝑡𝑖󸀠 means schedule arrival time, 𝑡𝑎𝑐𝑡𝑖 means
󸀠
of subsequent flights. On the other hand, we have to keep the actual departure time, 𝑡𝑎𝑐𝑡𝑖 means actual arrival time, 𝑡𝑏𝑢𝑓, 𝑡𝑠𝑒𝑟
scheduled departure time of the next flight unchanged and and 𝑡𝑡𝑢𝑟 means scheduled buffer time, standard ground service
reserve enough turnaround time (greater than the necessary time and scheduled turnaround time, respectively.
turnaround time) for the next flight. This means that we can One aircraft flies from airport 1 to airport 4, if airport 1
delay the scheduled departure time of the current flight, and belongs to one of the 10 busiest airports in the previous sta-
the maximum amount of delay is equal to the schedule buffer tistics, then we delay the scheduled departure time of flight 2
time between the current and the next flight operated by the from 𝑡2 to 𝑡𝑎𝑐𝑡2, and the amount of delay is equal to the sched-
same aircraft. According to our statistical results, the sched- uled turnaround time 𝑡𝑡𝑢𝑟3 between the flight 2 and the flight
uled turnaround time varies greatly between different flights 3 minus the necessary turnaround time 𝑇𝑡𝑢𝑟. All in all, if the
or different aircraft, but the required standard ground service time interval 𝛥t between the actual arrival time of flight 1 and
time is about 30–50 minutes, so we set the necessary turna- the schedule departure time of flight 2 is larger than required
round time 𝑇𝑡𝑢𝑟 to 30, 40, and 50 minutes as mentioned earlier. ground service time, flight 2 will take off on time.
Journal of Advanced Transportation 7

4. Empirical Results we eliminated effects of PF, the distribution of departure delay

exhibit an obvious PL feature instead of SPL. The queue model
In order to verify the effectiveness and practicability of our which executes the highest-priority item on its list helps to
method, we collect additional six-month data of flight opera- understand mechanism of PL feature. We consider that the
tion in the first half of 2018. We will use the method of this mechanism of SPL distribution of flight departure delay is the
article to adjust the flight planning and compare the number results of aircraft queue for take-off due to airports congestion
of delayed flights before and after adjustment for the first six and propagation delay caused by late-arriving aircraft.
months of 2018. From the previous statistical results, we know Based on the above mechanism, we develop a specific
that the 10 airports with the busiest and the highest delay ratio measure to mitigate propagation delay without increasing
are SFO, EWR, JFK, LGA, MIA, PBI, ORD, BOS, LAX and operational costs. Specifically, if one aircraft takes off from an
SEA. We assume that if one flight departs from one of these airport with a higher delay ratio, we delayed the schedule
10 airports, it will generate newly formed delay and cause departure time of the next flight operated by the same aircraft,
another flight immediately after it with the same aircraft also which is equal to the schedule buffer time between the next
to delay. However, strictly speaking, the latter flight delay may flight and the subsequent flight. It is proved that our approach
be not merely attributed by a late arrival of the flight immedi- is pretty effective in reducing flight delay, although it is not
ately preceding it, but also be attributed by one or more of significant for flights with larger delay.
other factors. In other words, sometimes the actual departure In addition, our approach is based on the predictability
delay is hard to predict when we change the flight plan by our of propagation delays and mathematical induction, which
method, while the delay only caused by PF is not. Therefore, provides a new way to optimize flight schedules. Although
to simplify the prediction of current flight delays in the present this is by no means intended as a exhaustive study, it
work, we do not take into account the newly formed delay nonetheless provides a starting point to motivate future
when the last flight by the same aircraft departed from one of research, which is more accurate forecasting of the newly
the 10 highest delay ratio airports. formed delays and finding the optimal amount of slack that
The six-month data comprehends 463322 flight operation we redistributed.
records, and a total of 84828 flights departing from the 10
highest delay ratio airports. Actually, since there are typically
no flights between 0 and 6 o’clock, delay on the last flight of Data Availability
each day does not propagate to the first flight of the next day.
Therefore, without considering the delayed propagation of the The data used to support the findings of this study can be
last flight per day, we only adjust the schedule departure time found from the website of the Bureau of Transportation
for 72902 flights instead of 84828 flights. Comparing the Statistics (BTS) at http://www.bts.gov.
results before and after adjustment, we find that the departure
delay ratio dropped from 13.91% to 12.06%, 12.25% and
12.39% with 𝑇𝑡𝑢𝑟 equal to 30 minutes, 40 minutes and 50 min- Conflicts of Interest
utes, respectively. The change in the number of delayed flights
in each delay interval is presented in Figure 6. The authors declare that there is no conflict of interest regard-
Obviously, we can see that the number of delayed flights ing the publication of this paper.
in almost all delay intervals has decreased. And the smaller
the necessary turnaround time 𝑇𝑡𝑢𝑟, the more the delay and
delay ratio will be reduced. But we cannot set 𝑇𝑡𝑢𝑟 too small in
Acknowledgments
our method, because large aircrafts require a relatively long This work is partially supported by National Nature Science
turnaround time, small 𝑇𝑡𝑢𝑟 does not correspond to actual. The Foundation of China under Grant Nos. 11775111, 61773203,
other reason is that the operation of the flights is full of many and U1833126. And we thank Dr. Bo Zhou for revising our
uncertain factors, the slack time is reserved to help deal with manuscript.
some unexpected situations and improve the robustness of the
flight plan. On the other hand, our method is pretty effective
in the case of short delay, but not in the case of long delay. This References
is due to the limited slack time reserved by the airline in for-
mulating flight plan. Many delayed flights with small delay [1] E. B. Peterson, K. Neels, N. Barczi, and T. Graham, “The
have been able to take off on time after our measure, but flights economic cost of airline flight delay,” Journal of Transport
with larger delay have only slightly reduced delay. Economics and Policy (JTEP), vol. 47, no. 1, pp. 107–121, 2013.
[2] G. Santos and M. Robin, “Determinants of delays at European
airports,” Transportation Research Part B: Methodological,
5. Conclusion vol. 44, no. 3, pp. 392–403, 2010.
[3] S. AhmadBeygi, A. Cohn, and M. Lapp, “Decreasing airline
By data mining and statistical analysis, we study the distribu- delay propagation by re-allocating scheduled slack,” SSRN
tion characteristics and inherent mechanism of flight depar- Electronic Journal, 2008.
ture delay for DL. From the statistical results, we find that the [4] K.F. Abdelghany, S.S. Shah, S. Raina, and A.F. Abdelghany, “A
distribution of flight departure delay follows SPL, and when model for projecting flight delays during irregular operation
8 Journal of Advanced Transportation

conditions,” Journal of Air Transport Management, vol. 10, no. 6, Research Part E: Logistics and Transportation Review, vol. 66,
pp. 385–394, 2004. pp. 103–114, 2014.
  [5] S. AhmadBeygi, A. Cohn, and Y. Guan, Analysis of the Potential [23] N. Kafle and B. Zou, “Modeling flight delay propagation: a new
for Delay Propagation in Passenger Aviation Flight Networks, analytical-econometric approach,” Transportation Research Part
University of Michigan, 2007. B: Methodological, vol. 93, pp. 520–542, 2016.
  [6] S. Lan, J.-P. Clarke, and C. Barnhart, “Planning for robust airline [24] N. Pyrgiotis, K. M. Malone, and A. Odoni, “Modelling delay
operations: optimizing aircraft routings and flight departure propagation within an airport network,” Transportation
times to minimize passenger disruptions,” Transportation Research Part C: Emerging Technologies, vol. 27, pp. 60–75, 2013.
Science, vol. 40, no. 1, pp. 15–28, 2006. [25] J.-T. Wong and S.-C. Tsai, “A survival model for flight delay
 [7]  P. T. R. Wang, L. A. Schaefer, and L. A. Wojcik, “Flight propagation,” Journal of Air Transport Management, vol. 23,
connections and their impacts on delay propagation,” in The pp. 5–11, 2012.
22nd Digital Avionics Systems Conference, 2003. DASC’03, [26] G. Stojković, F. Soumis, J. Desrosiers, and M. M. Solomon, “An
pp. 1–9, IEEE, 2003, 1: 5. B. 4–5, vol. 1. optimization model for a real-time flight scheduling problem,”
  [8] L. Ionescu, C. Gwiggner, and N. Kliewer, “Data analysis of Transportation Research Part A: Policy and Practice, vol. 36,
delays in airline networks,” Business & Information Systems no. 9, pp. 779–788, 2002.
Engineering, vol. 58, no. 2, pp. 119–133, 2016. [27] B. Rexing, C. Barnhart, T. Kniker, A. Jarrah, and
  [9] H. Fricke and M. Schultz, “Delay impacts onto turn around N.  Krishnamurthy, “Airline fleet assignment with time
performance,” Optimal Time Buffering for Minimizing Delay windows,” Transportation Science, vol. 34, no. 1, pp. 1–20, 2000.
Propagation, ATM Seminar, 2009. [28] M. Stojković and F. Soumis, “An optimization model for the
[10] P. Fleurquin, J. J. Ramasco, and V. M. Eguiluz, “Systemic delay simultaneous operational flight and pilot scheduling problem,”
propagation in the US airport network,” Scientific Reports, Management Science, vol. 47, no. 9, pp. 1290–1305, 2001.
vol.  3, no. 1, p. 1159, 2013. [29] http://www.bts.gov.
[11] C. Li-Ping, W. Ru, S. Hang et al., “Structural properties of US flight [30] N. Xu, G. Donohue, K. B. Laskey, and C.H. Chen, “Estimation
network,” Chinese Physics Letters, vol. 20, no. 8, pp. 1393–1396, of Delay Propagation in the National Aviation System using
2003. Bayesian Networks,” 6th USA/Europe Air Traffic Management
[12] W. Ru and C. Xu, “Hierarchical structure, disassortativity and Research and Development Seminar, FAA and Eurocontrol,
information measures of the US flight network,” Chinese Physics Baltimore, MD, 2005.
Letters, vol. 22, no. 10, pp. 2715–2718, 2005. [31] C.-L. Wu and R. E. Caves, “Aircraft operational costs and
[13] J. Lin and Y. Ban, “The evolving network structure of US airline turnaround efficiency at airports,” Journal of Air Transport
system during 1990–2010,” Physica A: Statistical Mechanics and Management, vol. 6, no. 4, pp. 201–208, 2000.
its Applications, vol. 410, pp. 302–312, 2014. [32] H. Idris, J.-P. Clarke, R. Bhuva, and L. Kang, “Queuing model
[14] A. L. Barabási, R. Albert, and H. Jeong, “Scale-free characteristics for taxi-out time estimation,” Air Traffic Control Quarterly, vol.
of random networks: the topology of the world-wide web,” 10, no. 1, pp. 1–22, 2002.
Physica A: Statistical Mechanics and its Applications, vol. 281, [33] A. Cobham, “Priority assignment in waiting line problems,”
no. 1–4, pp. 69–77, 2000. Journal of the Operations Research Society of America, vol. 2,
[15] R. Albert and A.-L. Barabási, “Statistical mechanics of complex no. 1, pp. 70–76, 1954.
networks,” Reviews of Modern Physics, vol. 74, no. 1, pp. 47–97, [34] P. G. Ipeirotis, “Analyzing the amazon mechanical turk
2002. marketplace,” XRDS: Crossroads, The ACM Magazine for
[16] A.-L. Barabási and R. Albert, “Emergence of scaling in random Students, vol. 17, no. 2, pp. 16–21, 2010.
networks,” Science, vol. 286, no. 5439, pp. 509–512, 1999. [35] Y. J. Wang, Y. K. Cao, C. P. Zhu et al., “Characterizing departure
[17] A.-L. Barabási, “The origin of bursts and heavy tails in human delays of flights in passenger aviation network of United States,”
dynamics,” Nature, vol. 435, no. 7039, pp. 207–211, 2005. arXiv preprint, 2017, https://arxiv.org/abs/1701.05556.
[18] R. Albert and A.-L. Barabási, “Topology of evolving networks: [36] J. H. Gittell, “Cost/quality trade-offs in the departure process?
local events and universality,” Physical Review Letters, vol. 85, Evidence from the major US airlines,” Transportation Research
no. 24, pp. 5234–5237, 2000. Record, vol. 1480, 1995.
[19] H. Chang, B.-B. Su, Y.-P. Zhou, and D.-R. He, “Assortativity
and act degree distribution of some collaboration networks,”
Physica A: Statistical Mechanics and its Applications, vol. 383,
no. 2, pp. 687–702, 2007.
[20] C.-H. Fu, Z.-P. Zhang, H. Chang et al., “A kind of collaboration–
competition networks,” Physica A: Statistical Mechanics and its
Applications, vol. 387, no. 5-6, pp. 1411–1420, 2008.
[21] Y.-L. Wang, T. Zhou, J.-J. Shi, J. Wang, and D.-R. He, “Empirical
analysis of dependence between stations in Chinese railway
network,” Physica A: Statistical Mechanics and its Applications,
vol. 388, no. 14, pp. 2949–2955, 2009.
[22] P. Baumgarten, R. Malina, and A. Lange, “The impact of hubbing
concentration on flight delays within airline networks: an
empirical analysis of the US domestic market,” Transportation
 International Journal of

Rotating Advances in
Machinery Multimedia

The Scientific
Engineering
Journal of
Journal of

Hindawi
World Journal
Hindawi Publishing Corporation Hindawi
Sensors
Hindawi Hindawi
www.hindawi.com Volume 2018 http://www.hindawi.com
www.hindawi.com Volume 2018
2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal of

Control Science
and Engineering

Advances in
Civil Engineering
Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Submit your manuscripts at

www.hindawi.com

Journal of
Journal of Electrical and Computer
Robotics
Hindawi
Engineering
Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

VLSI Design
Advances in
OptoElectronics
International Journal of

International Journal of
Modelling &
Simulation
Aerospace
Hindawi Volume 2018
Navigation and
Observation
Hindawi
www.hindawi.com Volume 2018
in Engineering
Hindawi
www.hindawi.com Volume 2018
Engineering
Hindawi
www.hindawi.com Volume 2018
Hindawi
www.hindawi.com www.hindawi.com Volume 2018

International Journal of
International Journal of Antennas and Active and Passive Advances in
Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration
Hindawi Hindawi Hindawi Hindawi Hindawi
www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018