Agent-based Modelling of Aircraft Boarding Methods
Serter Iyigunlu, Clinton Fookes and Prasad Yarlagadda
Science and Engineering Faculty, Queensland University of Technology, Brisbane, QLD 4000, Australia
Keywords: Agent-based Modelling, Aircraft Boarding, Airport Operations, Simulation.
Abstract: We implemented six different boarding strategies (Wilma, Steffen, Reverse Pyramid, Random, Blocks and
By letter) in order to investigate boarding times for Boeing 777 and Airbus 380 aircraft. We also introduce
three new boarding methods to find the optimum boarding strategy. Our models explicitly simulate the
behaviour of groups of people travelling together and we explicitly simulate the timing to store their
luggage as part of the boarding process. Results from the simulation demonstrates the Reverse Pyramid
method is the best boarding method for Boeing 777, and the Steffen method is the best boarding method for
Airbus 380. For the new suggested boarding methods, aisle first boarding method is the best boarding
strategy for Boeing 777 and row arrangement method is the best boarding strategy for Airbus 380. Overall
best boarding strategy is aisle first boarding method for Boeing 777 and Steffen method for Airbus 380.
1 INTRODUCTION
Airlines only start to earn profits when their aircraft
commence flying. In order to increase profit,
turnaround time should be minimised. Airplane
turnaround time is calculated between airplane’s
arrival and departure times (Briel et al., 2005; Briel
et al., 2005). Factors contributing to turnaround time
usually divided into seven groups which are
disembarkation, baggage unloading, refuelling the
aircraft, cargo unloading, aircraft maintenance,
cargo and baggage loading and passenger boarding
(Soolaki et al., 2012). In this study, only the
passenger boarding process of turnaround time
investigated, as this is a significant contributor to
delays in turnaround time and it is a source of much
uncertainty without clear options for optimisation.
Figure 1.1 explains aircraft turnaround time.
Aircraft boarding methods have been
investigated since the 1980’s, and in particular
simulation models are becoming prevalent because
they are cost-effective, convenient, risk-free, and
easy to change. The boarding of passenger aircraft
has been a problem since the beginning of airline
industry (Soolaki et al., 2012).
In this study, a simulation model of two different
aircraft types (Boeing 777-200 and Airbus 380) are
constructed and used to examine six different
aircraft boarding strategies using an agent-based
modelling apporoach in Anylogic simulation
program. The outline of the rest of the paper is as
follows. Section 2 presents the background literature
on Aircraft boarding methods literature
search,
Section 3 represents existing and new aircraft
boarding strategies, Section 4 represents discussion
about boarding strategies and Section 5 represents
conclusion.
2 AIRCRAFT BOARDING
METHODS
Aircraft boarding represents a major source of
delays and is a direct result of how smooth
passengers flow, engage and interact in the boarding
process. Passengers enter the aircraft one by one,
search for their seats, place their luggage into the bin
above their seats and sit down. Most airlines utilise
assigned seating, i.e. passengers cannot change
their seat numbers if they are printed their boarding
tickets (Cimler et al., 2006).
Marelli et al. (1998) suggested a new passenger
embarking/disembarking simulation model to
indicate different boarding methods and different
arrangements on a Boeing 757 plane. This method
uses steady variables such as the velocity of
passengers. Van Landeghem and Beuselinck (2002)
analysed different aircraft boarding designs and they
investigated only short haul flights and aircrafts
148
Iyigunlu S., Fookes C. and Yarlagadda P..
Agent-based Modelling of Aircraft Boarding Methods.
DOI: 10.5220/0005033601480154
In Proceedings of the 4th International Conference on Simulation and Modeling Methodologies, Technologies and Applications (SIMULTECH-2014),
pages 148-154
ISBN: 978-989-758-038-3
Copyright
c
 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 2.1: Aircraft turnaround time explanation (Landeghem and Beuselinck, 2002).
typically have 80 to 150 seats. Van den Briel et al.
(2003, 2005) applied programming, noticed data and
simulation to study the aircraft boarding strategy.
Ferrari and Nagel (2006) used different
arrangements to assess different aircraft boarding
methods and suggested a new aircraft boarding
method that contains boarding groups, it consists of
different types of schemes such as early and/or late
passenger but did not cover aisle interferences.
Bazargan (2007) evaluated the meddling between
the passengers that cause waiting during the
boarding process and built a new integer
programming model to reduce the interferences, the
suggested system is good for reducing boarding time
but it only includes single aisle aircraft types.
Nyquist and McFadden (2008) demonstrated the
cost-effective way to board passengers including
carry-on luggage and boarding trough two doors.
Steffen (2008) implemented the Markow Chain
Monte Carlo algorithm and numerical simulation to
research the aircraft boarding method in order to
minimise boarding time. Steiner and Philipp (2009)
sought some actions like introducing pre-boarding
areas that can reduce the boarding time and turn
time; they investigate only two different boarding
strategies (Back to Front method and Random
method). Tang et al. (2012) used the pedestrian flow
theory to suggest an aircraft boarding method to
research the random boarding strategy (Tang et al.,
2012).
The literature introduced different boarding
strategies in order to reduce boarding time. These
include:
1.) Wilma Method: Passengers seated at the
windows boarding first, followed by the middle and
aisle seats. Inside the group passengers are ordered
randomly, therefore there are no seat interferences.
2.) Reverse Pyramid (back to front): passengers
board from the rear rows first and then boarding
gradually moves forward.
3.) Steffen method: Neighbouring passengers in
line are sitting two rows apart from each other in
equivalent seats. (e.g., 12A, 10A, 8A, 6A. etc.)
Airplane turn
time
30-60 minutes
Passengers leave
plane (Deplane)
10-15 minutes
Aircraft cleaning
10-15 minutes
Passenger boarding
(Enplane)
10-30 minutes
Call for passengers
Passenger missing
Transfer passengers'
late arrival
Boarding pass
control at gate
entrance
Card reader jammed
Handluggage
retrieval at gate
Passenger installation
in aircraft
Handluggage storage
insufficient
Seat assignment
errors
Seats occupies out of
sequence
Agent-basedModellingofAircraftBoardingMethods
149

(Steffen and Hotchkiss, Experimental test of airplane
boarding methods, 2012).
4.) Random method: All passengers are boarding
together, without particular arrangement (Cimler et
al., 2009).
5.) Blocks: Boarding in blocks. The back rows
are first boarded, then the front blocks boarded and
finishing with the centre rows blocks (Cimler et al.,
2009).
6.) By letter: This is a special type of boarding
method, in which each Class contains all seats with
the same tag. (A to F) (H. Van Landeghem, 2002)
Generally, the actual boarding process comprises
three steps (Landeghem and Beuselinck, 2002):
1. The gate agent announced beginning of the
boarding process. Passengers start queuing at the
gate. There might be delay because of the late arrival
passengers. (Call- Off System)
2. The gate agent checks the boarding pass of
each passenger, and records their entry by using a
ticket reader. This step in the procedure is the only
moment we have control on who enters at what time.
3. Finally passengers enter the airplane through
the bridge and the front door of the aircraft.
There are two types of interferences in the
boarding process. One of them is seat interference
and the other one is aisle interferences (Briel et al.,
The Aircraft Boarding Problem, 2003). By
minimizing the total expected number of seat and
aisle interferences is the best way to minimize
boarding time.
Seat interference happens when the middle
and/or aisle seat passenger boards earlier than a
window or middle seat passenger that sits on the
same side and row of the aircraft. For instance, a
passenger is seated in seat 12C. When the passenger
with seat 12B or 12A boards the aircraft, passenger
13C must leave their seat in order to give way to
passenger 13B or 13A.
In definition of aisle interference is once a
passenger boarding the aircraft must wait for the
passenger in front of them to go their seat and to
stow their luggage before carrying on to their seat.
Aisle interference can happen within one group, or
between two ensuing groups (Briel et al., 2003).
Figure 2.1 illustrates both seat and aisle
interferences.
Figure 2.2: Seat (above) and aisle (below) interferences
illustration (Junior et al., 2008).
3 AGENT-BASED MODELLING
Agent-based modelling offers a way to model social
systems that are composed of agents who interact
with influence each other. Agent-based modelling is
a way to model the dynamics of complex systems
and complex adaptive systems. The essential
characteristic of an agent is the ability to make
independent decisions (Macal and North, 2010).
An agent should have the following
characteristics:
 An agent is a self-contained, modular and
uniquely identifiable individual.
 An agent is autonomous and self-directed.
 An agent has a state that varies over time.
 An agent is social having dynamic
interactions with other agents that influence
its behaviour (Macal and North, 2010).
In addition to the above statements, the agent
should have the following characteristics:
 An agent should be a reactive that responses
to changes in its environment.
 An agent is a proactive.
 An agent should be adaptable and has many
ways of obtaining targets (Ronald et al.,
2007).
3A 3B 3C 3D 3E 3F
4A
4B
4C 4D
4E
4F
5A 5B 5C 5D 5E 5F
3A 3B 3C 3D 3E 3F
4A 4B 4C 4D 4E 4F
5A 5B 5C 5D 5E 5F
6A 6B 6C 6D 6E 6F
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
150

4 AIRCRAFT TYPES AND
IMPLEMANTATION OF
BOARDING STRATEGIES
In this study, Boeing 777 and Airbus A380 aircrafts
were used because of their seat capacity and widely
used for international travelling. Boeing 777 has
two aisles, total 42 rows and its capacity is 313
passengers. Airbus 380 is consisted of two aisles,
total of 48 rows and its capacity is 498 passengers.
Airbus 380 is also a double-deck aircraft.
4.1 Boarding Strategies
In practice most commercial airline companies use
what is called a back-to-front boarding strategy. The
idea is to free the aisle of congestions as much as
possible (Audenaert et al., 2009). The back to front
approach results in the bottleneck being created in a
reduced area of the aisle between passengers of the
same group to store their carry-on luggage and to
reach their assigned seat in an appropriate way
(Briel et al., The Aircraft Boarding Problem, 2003).
The distribution of passengers travelling through
the air transportation system includes those
travelling for business purposes and those travelling
for tourism and leisure. Business travellers
predominantly travel alone; however, those
travelling for tourism and leisure are regularly in
groups. To reflect this reality in the simulation, all
passengers are randomly drawn from a distribution
of group sizes where the group size predominantly
varies from one to five.
In the arrival, passengers are assigned an
attribute in order to determine their seats and
behaviour in the group.
Table 4.1: Passengers’ attributes represent baggage
waiting time, golden passengers percentage, children
percentage and group size.
SimulationAttributes
Baggageinflight
waitingtime
Minimum5seconds,
average25seconds
andmaximum60
seconds
Goldenpassengers
probability
0.5%
Passengerswith
childprobability
1%
Groupsize Varyingfrom1to5

If a passenger has a luggage in flight, there is a
delay in their row, which is minimum 5 second,
average 25 seconds and maximum 60 seconds. Table
3.1 shows passengers’ attributes in the simulation
program. In the beginning of the simulation, these
attributes are applied to the passengers. Golden
passengers are assigned at the rate of 0.5 % of all
passengers. Child passengers are assigned according
to the probability of 1 %. 1% of all passengers are
children in the simulation program.
Figure 4.1: Aircraft boarding simulation process.
Table 4.2: Boarding times overall.
Each boarding method was run ten times in order
to investigate boarding methods. The reverse
pyramid method is the fastest boarding method for
the Boeing 777-200 as it can be seen in Table 3.2.
The reverse pyramid method however is not the
fastest boarding method for Airbus 380. It is because
difference between aircraft lay outs. Airbus 380 is a
two-level aircraft and its best efficient boarding
method is the Steffen method.
4.2 Proposed New Boarding Methods
In this section, the new boarding strategies will be
investigating in order to reduce boarding times and
turnaround time. The proposed new boarding
methods are determined by testing all boarding
PassengersArriveat
thegate
PassengersGoto
WaitingAreafor
Boarding
FirstBoardingGroup:
Passengerswithchild
SecondBoarding
Group:Groupsizeof
5peopleormore
ThirdBoarding
Group:Gold
members
FourthBoarding
Group:FirstClass
Passengers
FifthBoardingGroup:
BusinessClass
Passengers
EconomyClass
Passengersare
Boardedaccordingto
theriboardingzone
Boarding
Time(minutes)
StandardDeviation
Boarding
Time(minutes)
StandardDeviation
WilmaMethod 31.4487 2.041797686 55.59
8
1.943370897
Reverse Pyramid
(Back To Front)
29.6414 1.896140536 43.988 1.022474884
SteffenMethod 40.3306 1.518928365 39.463
2
1.296724489
RandomMethod 30.6855 2.40296924 51.6033 1.20908018
8
Bl ocks 35.0046 2.009044449 48.6333 1.474254467
ByLetter 47.8871 2.058640946 64.0214 2.793316873
A380B777
PassengerNumber=313 PassengerNumber=498
Agent-basedModellingofAircraftBoardingMethods
151

alternatives. The following figures show the new
boarding strategies and aisle delays for each
boarding strategy.
Figure 4.2: Aisle boarding first method illustrations. Seat
type 1 enters the aircraft first, seat type 2 enters the aircraft
second and seat type 3 enters the aircraft third.
Figure 3.2 shows the aisle boarding first method.
In this boarding method, all middle aisles of the
aircraft are boarded first (Seat type1), following
window seats (Seat type2), middle seats (Seat type2)
and aisle passengers (Seat type3) are boarded
respectively.
Figure 4.3: Aisle boarding first method comparison of
aisle delay times for two different aircraft types (Boeing
777 and Airbus 380).
Figure 3.3 shows aisle delay times while
boarding in the Boeing 777 and Airbus 380,
respectively. X axis represents waiting times in
minutes scale and Y axis represents the percentage
of passengers. As seen in the graph, Boeing 777
aisle delay time is much smaller than Airbus 380
because of seat capacity.
Figure 4.4: Row arrangement method illustrations. Seat
type 1 enters the aircraft first, seat type 2 enters the aircraft
second and seat type 3 enter the aircraft third.
Figure 3.4 shows the row arrangement method. In
this boarding method, passengers enter the aircraft
according to their seat numbers. For example, row
12, row 15, row 18 etc., will be boarded first (Seat
type1), row 13, row 16, row 19 etc., will be boarded
second (Seat type2), row 14, row 17, row 20 etc.,
will be boarded third (Seat type3).
Figure 3.5 shows aisle delay time for Boeing 777
and Airbus 380 aircrafts respectively for the row
arrangement boarding method. Boeing 777 aisle
delay time is slightly smaller than Airbus 380.
Figure 4.5: Row arrangement method comparison of aisle
delay times for two different aircraft types (Boeing 777
and Airbus 380).
Figure 3.6 explains boarding method 3.
Figure 4.6: Row blocks method illustrations. Seat type 1
enters the aircraft first, seat type 2 enters the aircraft
second, seat type 3 enters the aircraft third, seat type 4
enters the aircraft fourth, seat type 5 enters the aircraft
fifth and seat type 6 enters the aircraft sixth.
Figure 3.7 represents the comparison of aisle
delay times for the Boeing 777 and Airbus 380
aircraft for the row blocks boarding method. As seen
in the figure, Boeing 777 aisle delay time is much
smaller than Airbus 380 delay time. On Airbus 380,
aisle delay time is approaching 20 minutes but the
average delay time is 11 minutes.
Seat type 1 Seat type 2 Seat type 3 Seat type 4
Seat type 1 Seat type 2 Seat type 3
Seat type 1 Seat type 2 Seat type 3 Seat type 4
Seat type 5 Seat type 6
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
152

Figure 4.7: Row blocks method comparison of aisle delay
times for two different aircraft types (Boeing 777 and
Airbus 380).
Table 4.3: Boarding times overall.
Boarding
Time(minutes)
Standard
Deviation
Boarding
Time(minutes)
Standard
Deviation
B777 A380
Passenger number=313 Passenger number=498
Aisle
boarding
first method
26.05
1.5510 46.66 1.3224
Row
arrangement
method
34.505 1.3243
42.773
1.3764
Row blocks
method
38.83 1.0909 61.627 1.4536
As seen in the Table 3.3, Aisle boarding first
method is the fastest method for Boeing 777 but for
Airbus 380
row arrangement method is given better
boarding result.
5 DISCUSSION
Boarding methods play a vital role in order to reduce
the turnaround time and increase airlines’ efficiency.
Airlines earn profit when their aircraft fly. In order
to increase the number of flying aircraft each day,
the turnaround time must be decreased.
The turn time annual cost is found from the
following expression (Nyquist & McFadden, 2008):
        365
where;
C: annual cost;
B = average boarding time (in minutes);
M = cost per minute on the ground;
D = average number of daily flights.
For example, if an airline spends $40 dollars on
the ground per minute, and has 500 flights daily,
Table 4.1 shows the annual costs for each boarding
method.
Table 5.1: Annual costs for each boarding method.
Boarding
time(minutes)
Annual cost
for the
Boeing 777
Boarding
time
(minutes)
Annual
Cost for
the Airbus
380
B777 A380
Passenger number=333 Passenger number=498
Aisle
boarding first
method
26.05 $190,165.00
0
46.66 $340,618.
000
Row
arrangement
method
34.505 $251,886.50
0
42.773 $312,242.
900
Row blocks
method
38.83 $283,459.00
0
61.627 $449,877.
100
As seen in Table 4.1, reducing boarding times
can play a significant role for airlines to make profit.
For the Boeing 777 aircraft boarding methods, the
difference between aisle boarding first method and
row blocks method is $93,294.000. An airline can
save $93,294.000 each year only by using aisle
boarding first boarding strategy.
6 CONCLUSION
In this study, we have implemented existing
passenger boarding methods and three new boarding
methods on Boeing 777 and Airbus 380 aircrafts.
The reverse pyramid method is the fastest boarding
method for Boeing 777 and Steffen boarding method
is the fastest method for Airbus 380 out of the
existing methods.
Aisle boarding first method is the fastest
boarding method for Boeing 777 because of the
reduced aisle interferences and row arrangement
method is the fastest boarding method for Airbus
380 due to increased smoothness of passenger flow.
Aisle boarding first method is the fastest method
for the Boeing 777 type of aircraft and Steffen
boarding method is the fastest boarding method for
Airbus 380 type of aircrafts.
It is recommended that future research be
undertaken in the following areas:
1. Investigate long haul flights
Nowadays, international flights are playing
important roles in airline industry. Future research
should concentrate on long haul flights rather than
short haul flights and investigate new type aircrafts
for boarding methods such as Boeing 777
Dreamliner.
2. Investigate pre-boarding zones
Future research should investigate pre-boarding
zones and introduce boarding zone announcement
system via mobile phones. Passengers should know
where and when their boarding zones will be
boarded.
Agent-basedModellingofAircraftBoardingMethods
153

REFERENCES
Audenaert, J., Verbeeck, K., & Berghe, G. V. (2009).
Multi-Agent Based Simulation for Boarding. The 21st
Belgian-Netherlands Conference on Artificial
Intelligence (pp. 3-10). Belgium: BNAIC.
Bandini, S., Manzoni, S., & Vizzari, G. (2012). Agent
Based Modeling and Simulation. Computer Science,
105-117. doi:10.1007/978-1-4614-1800-9_7
Bauer, D., Seer, S., & Brändle, N. (2007). Macroscopic
Pedestrian Flow Simulation for Designing Crowd
Control Measures in Public Transport after Special
Events. SCSC 2007, (pp. 1035-1042).
Bazargan, M. (2007). A linear programming approach for
aircraft boarding strategy. European Journal of
Operational Research, 394-411.
Briel, M. H., Villalobos, J. R., & Hogg, G. L. (2003). The
Aircraft Boarding Problem. In Proceedings of the 12th
Industrial Engineering Reseach Conference(CD-
ROM). IERC.
Briel, M. H., Villalobos, J. R., Hogg, G. L., Lindemann,
T., & Mule, A. V. (2005). America West Airlines
Develops Efficient Boarding Strategies. ProQuest, 35,
191-201.
Cimler, R., Kautzka, E., Olsevicova, K., & Gavalec, M.
(2009). Agent-based model for comparison of aircraft
boarding methods. Proceedings of 30th International
Conference Mathematical Methods in Economics, (pp.
73-78).
Ferrari, P., & Nagel, K. (2006). Robuttnes of Efficient
Passenger Boarding Strategies for Airplane.
Transportation Research Record:Journal of the
Transportation Research Board, 44-54.
Harney, D. (2002). Pedestrian modelling:Current methods
and future directions. Road and Transport Research,
38-48.
Junior, E. C., Silva, J. L., Briel, M. H., & Villalobos, J. R.
(2008). Aircraft Boarding Fine-Tuning. XIV
International Conference on Industrial Engineering
and Operations Management. ICIEOM.
Klügl, F., & Bazzan, A. L. (2012). Agent-Based Modeling
and Simulation. AI Magazine, 29-40.
Landeghem, H. V., & Beuselinck, A. (2002). Reducing
passenger boarding time in airplane: A simulation
based approach. European Journal of Operational
Research, 294-308.
Ma, W., Kleinschmidt, T., Fookes, C., & Yarlagadda, P.
K. (2011). Check-in processing:Simulation of
passengers with advanced traits. In S. Jain, R. R.
Creasey, J. Himmelspach, K. P. White, & M. Fu (Ed.),
Proceedings of the 2011 Winter Simulation
Conference (pp. 1783-1794). IEEE.
Macal, C. M., & North, M. J. (2010). Toward Teaching
Agent-Based Simulation. Proceedings of the 2010
Winter Simulation Conference (pp. 268-277).
Baltimore: IEEE.
Marelli, S., Mattocks, G., & Merry, R. (1998). The Role of
Computer Simulation in Reducing Airplane Turn
Time. Aero Magazine.
Nyquist, D. C., & McFadden, K. L. (2008). A study of the
airline boarding problem. Journal of Air Transport
Management, 197-204.
Papadimitriou, E., Yannis, G., & Golias, G. (2009). A
critical assessment of pedestrian behaviour models.
Transportation Research, 242-255.
Qiu, F., & Hu, X. (2010). Modeling group structures in
pedestrian crowd simulation. Simulation Modelling
Practice and Theory, 190-205.
Soolaki, M., Mahdavi, I., Mahdavi-Amiri, N.,
Hassanzadeh, R., & Aghajani, A. (2012). A new linear
programming approach and genetic algorithm for
solving airline boarding problem. Applied
Mathematical Modelling, 4069-4072.
Steffen, J. H. (2008). Optimal boarding method for airline
passengers. Journal of Air Transportation
Management, 146-150.
Steffen, J. H., & Hotchkiss, J. (2012). Experimental test of
airplane boarding methods. Journal of Air Transport
Management, 64-67.
Steiner, A., & Philipp, M. (2009). Speeding up the
airplane boarding process by using pre-boarding areas.
9th Swiss Transport Research Conference (pp. 1-30).
STRC.
Tang, T., Huang, H., & Shang, H. (2012). A new
pedestrian-following model for aircraft boarding and
numerical tests. Nonlinear Dyn, 437-443.
SIMULTECH2014-4thInternationalConferenceonSimulationandModelingMethodologies,Technologiesand
Applications
154