International Journal of

Environmental Research
and Public Health

Article
Multiobjective Gate Assignment Model Considering
Carbon Emissions
Yixuan Shan, Yuwei Shao, Qi Yuan and Yu Jiang *

College of Civil Aviation, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China
* Correspondence: jiangyu07@[Link]

Abstract: It has been a main concern for governments to reduce the carbon emission of the aviation
industry. The paper proposes a multiobjective gate assignment model that considers the carbon
emission at the airport surface to facilitate environmental-friendly airport construction. Three
objectives are considered in the model to reduce carbon emissions, including the proportion of flights
assigned to the contact gate, aircraft taxiing fuel consumption, and gate assignment robustness. In
order to achieve better performance on all objectives, a Non-dominated Sorting Genetic Algorithm-
II (NSGA-II) is used to obtain the optimal results. The operation data from a domestic airport is
deployed to validate the model. The optimal results of the gate assignment model are compared with
the original scheme. It indicates that the proposed model can effectively reduce carbon emissions.
The study can provide a strategy for gate assignment to reduce carbon emissions and improve the
management of the airport.

Keywords: aviation; gate assignment; carbon mission; multiobjective model; NSGA-II

1. Introduction
In recent years, the environmental effect of civil aviation operations has become an im-
portant issue in air traffic management, especially for hub airports [1]. The flight congestion
Citation: Shan, Y.; Shao, Y.; Yuan, Q.; and delay resulting from the rapid growth of air traffic demand leads to a surge in carbon
Jiang, Y. Multiobjective Gate emissions. It has been reported that the carbon emissions of civil aviation transportation
Assignment Model Considering account for approximately 2% of the total carbon emissions of human activities [2]. It is
Carbon Emissions. Int. J. Environ. Res. necessary to find an efficient way for airports to operate more sustainably [3]. Therefore,
Public Health 2023, 20, 3952. https:// how to reduce the carbon emissions of aircraft operations has become a crucial problem
[Link]/10.3390/ijerph20053952 to solve.
Academic Editors: Xiaofeng Xu,
The gate assignment of aircraft determines the taxiing distance and utilization of ground
Wei Zhang, Yibin Ao and Cong support equipment (GSE) at the airport surface, which are the main sources of aviation
Cheng emissions [4,5]. Most existing studies focusing on gate assignments promote service quality,
operation efficiency, or operation cost rather than decreasing carbon emissions [6–10]. To the
Received: 27 December 2022 best of the authors’ knowledge, the problem of gate assignment optimization with the aim
Revised: 15 February 2023
of optimizing carbon emissions has yet to be adequately addressed. How to reduce taxiing
Accepted: 19 February 2023
carbon emissions and the extra emissions caused by conflicts at the surface when assigning
Published: 23 February 2023
gates to aircraft is the focus of this paper.
The contribution of this paper is the establishment of a multiobjective gate assignment
model to maximize the proportion of flights assigned to the contact gate, minimize taxiing
Copyright: © 2023 by the authors.
fuel consumption, maximize assignment robustness, and reduce the carbon emissions of
Licensee MDPI, Basel, Switzerland. operation. The gate assignment model can facilitate the establishment of environmentally
This article is an open access article friendly airports.
distributed under the terms and The remainder of the paper is structured as follows. A review of the literature related
conditions of the Creative Commons to gate assignment models and associated algorithms is presented in Section 2. The multi-
Attribution (CC BY) license (https:// objective gate assignment model considering carbon emissions and the solving algorithms
[Link]/licenses/by/ are described in Section 3. In Section 4, the operation data of domestic airports are used to
4.0/). validate the model. Finally, the conclusions are highlighted in Section 5.

Int. J. Environ. Res. Public Health 2023, 20, 3952. [Link] [Link]
Int. J. Environ. Res. Public Health 2023, 20, 3952 2 of 12

2. Literature Review
2.1. Gate Assignment Models
Many researchers have focused on the gate assignment problem. Most research has
established single-objective models to optimize passenger walking distance or operation
efficiency [11–13]. Some models integrate multiple objectives to optimize the gate assign-
ment problem [6,8,14]. Single-objective optimization is further divided into passenger
satisfaction, assignment plan robustness, airline benefits, etc. To improve airport capacity,
Xiong et al. [11] developed a gate assignment model with the objective of minimizing the
aircraft taxiing distance for multi-runway airports. Genc et al. [15] proposed minimizing the
number of flights parked in remote boarding gates and solved the model using a composite
algorithm. Wang et al. [12] proposed a single-objective model with the lowest cost of gate
assignment. Li [13] built a model based on the actual business rules of the airport, with the
maximum flight rate assigned to the contact gate as the optimization objective.
In recent years, more research has considered multiple interests simultaneously with
multiobjective integrated optimization. Xue et al. [6] considered the fairness among airlines
and established a multiobjective model for reducing the gate cost and taxi fuel consumption
cost of airlines. Shen et al. [7] proposed establishing a gate pre-assignment model with the
gate matching degree as the objective and integrating the near-airport flight-to-bridge ratio
to carry out bi-objective optimization to maximize the gate utilization and travel satisfaction
of passengers. Jiang et al. [8] constructed a multiobjective optimization model for gate
assignment with the objectives of minimizing the probability of ramp conflicts, the walking
distance of passengers for changing gates, and the number of passengers assigned to remote
boarding gates. From a new perspective, Kumar et al. [16] considered three optimization
objectives: minimum cost, maximum operating revenue, and robustness of airports, and
proposed a multiobjective assignment model. Das [14] developed a multiobjective gate
assignment model with the objectives of minimizing the travel distance by passengers and
maximizing the airport stores passed by passengers on their route. From the perspective of
airport resource use efficiency, Zheng et al. [17] constructed a bi-objective model with the
optimization of maximizing the match between gate and aircraft type and minimizing the
number of used gates.
In the field of air transport, due to the lack of technological breakthroughs in carbon-
based fuel sources, a small number of scholars have begun to study ways to reduce carbon
emissions from the perspective of airport operation management [18–21]. Sznajderman
et al. [5] modeled several scenarios characterized according to gate assignment policy and
analyzed different management model influences on the gaseous emissions generated
by aircraft and ground support equipment altogether. Evertse [22] studied the problem
of airport real-time taxiing path planning with the aim of minimizing emissions and
used CPLEX software to solve the established model. Soltani et al. [23] introduced a
MILP model to optimally control taxiing operations in an airport with an emphasis on
conflict and tow-truck usage to minimize fuel consumption considering the concept of
fixed parking positions. In summary, in terms of airport emission reduction management,
the research is mainly focused on the study of aircraft taxiing [22–26], while there are few
studies on gate assignment modeling considering multiple factors. Therefore, based on the
theory of multiobjective programming, this paper established a gate assignment model and
designed an algorithm to solve it for improved airport operation efficiency and reduced
carbon emissions.

2.2. Algorithms
In recent years, domestic and foreign scholars have adopted various algorithms to
solve different optimization objectives and constraints. Yan et al. [10] solved the joint
optimal scheduling problem of flight taxiing path planning and parking space allocation
in airports by an adaptive differential evolution algorithm but did not consider the flight
landing rate and aircraft taxiing fuel consumption. Lu [27] restored the aircraft running in
the scene through the NSGA-II algorithm but did not consider the taxiing fuel consumption,
Int. J. Environ. Res. Public Health 2023, 20, 3952 3 of 12

only the taxiing time. Based on the CART algorithm, Tang et al. [28] compared and analyzed
the prediction accuracy of apron configuration characteristic variables before and after
introduction in predicting the departure time of aircraft. However, the algorithm set taxiing
as barrier-free taxiing without considering obstacles when taxiing. Li et al. [29] improved
the genetic algorithm based on the structural characteristics of the flight area to reduce
the fuel consumption of the aircraft in the airport flight area to achieve the lowest fuel
consumption of the aircraft but did not consider the taxiing conflict time. Gök et al. [30]
used a sum-element heuristic algorithm to solve the scheduling of airport turnover tasks
and the route arrangement of the ground crew but did not include the conflict time at
the intersection. Yin et al. [31] introduced a dictionary sequential optimization algorithm
to allocate tarmac and runway for taxiing traffic in busy airports and allocated flights in
sequence under the constraints of parking space availability and runway utilization but did
not account for aircraft taxiing fuel consumption or the flight landing rate. According to
the principle of priority, Wen et al. [32] introduced the gate and flight labeling function to
design a labeling algorithm that can solve the scheduling model, which provided a feasible
means for optimizing the automatic allocation of gates.
In summary, the optimization research on gate assignment mainly focuses on im-
proving the heuristic intelligent optimization algorithm [29,33]. The research focuses on
establishing the optimization goal of the gate assignment model, which leads to its simple
constraints, low correlation with the actual airport business rules, and lack of practical
guiding significance. Traditional genetic algorithms have the advantages of wide range,
high robustness, and strong global search ability, but they easily fall into local solutions,
and the convergence results are easy to repeat [34]. The nondominated sorting genetic
algorithm (NSGA) has greatly improved speed due to layering individual relationships. In
this paper, nondominated sorting genetic algorithm II (NSGA-II) [35], with the mutually
exclusive characteristics of the three objectives in the model, is applied to optimize the
gate assignment problem and obtain the nondominated set for multiobjective optimiza-
tion problems.

3. Methodology
3.1. Multiobjective Gate Assignment Model
The gate assignment can directly influence the taxiing route of the flights. To be more
specific, when the flight is assigned to the contact gate, it will reduce the taxiing distance
of the flight, thus reducing carbon emissions. We propose increasing the proportion of
flights assigned to the contact gate and reducing the taxiing fuel consumption of all flights
to limit carbon emissions. In addition, the robustness of gate assignment will also impose
an impact on carbon emissions. Increasing the gate assignment robustness can decrease the
conflict caused by gate re-assignment and the waiting time of flights, which can benefit
carbon emission optimization. Therefore, three objectives are considered when modeling
the gate assignment: the proportion of flights assigned to the contact gate, the taxiing fuel
consumption, and the gate assignment robustness. The reduction in carbon emissions when
allocating the gates to flights can facilitate environmentally friendly airport construction.
The objectives of the model are as follows:
(1) Improve the proportion of flights assigned to the contact gate
The proportion of flights assigned to the contact gate refers to the proportion of flights
connecting with the terminal through the bridge. When the flight is assigned to a contact
gate, the taxiing distance of the flight will significantly decrease, which can help reduce
the carbon emissions of the flight. In addition, it has become a trend for bridges to be
equipped with auxiliary power unit (APU) replacement facilities. The APU replacement
facilities equipped in the gate are powered by electricity and thus do not generate carbon
dioxide and other gases. Improving the proportion of flights assigned to contact gates can
effectively reduce the carbon emissions of flights and improve the satisfaction of passengers
Int. J. Environ. Res. Public Health 2023, 20, 3952 4 of 12

because they can walk a shorter distance to board. The objective function (1) indicates the
maximization of the proportion of flights assigned to the contact gate.

Xik
Z1 = max ∑ ∑ (1)
i∈ F k∈G
N

where i is the number of flight pairs; F is the flight set; G is the gate set; Xik is the decision
variable, which is equal to one if the flight i is assigned to the gate k and zero otherwise;
and N is the number of flights.
(2) Reduce aircraft taxiing fuel consumption
The taxiing fuel consumption of flights is directly related to carbon emissions. There-
fore, a decrease in taxiing fuel consumption can facilitate environmentally friendly airport
construction. Moreover, it can also reduce the operation cost of airlines. In this paper,
taxiing fuel consumption is calculated for the surface taxiing process by weighting the
taxiing distance with the fuel flow rate when the aero-engine is idling-rated. The taxiing
fuel consumption of flight pairs is proportional to the aircraft engine fuel flow rate, taxiing
time and the number of engines [25]. The fuel consumption calculation formula of the
flight pair is shown in Equation (2).
dep
diarr + di
Fi = × ni × f i (2)
v
dep
where diarr is the taxiing distance of the arrival flight; di is the taxiing distance of the
departure flight; v is the average flight taxiing speed; ni is the number of engines; and f i is
the engine fuel flow rate.
Therefore, the objective function (3) indicates the minimization of fuel consumption.
dep
! !
dkarr + dk
Z2 = min ∑ ∑ Xik × ni × f i (3)
i∈ F k∈G
v

(3) Increase gate assignment robustness

Increasing the gate assignment robustness considering the conflict time, can improve
the utilization of gates, mitigate flight delays, and reduce the carbon emissions caused
by congestion. Therefore, how to measure the robustness of gate assignments needs
to be studied. The robustness of gate assignments refers to the anti-interference ability
of the assignment scheme when a flight delay occurs, which can be measured by the
expected conflict time. The smaller the expected conflict time is, the better the robustness
performance of the gate assignment schedule. Therefore, this paper extracts the actual
flight operation data of the airport to fit the flight delay distribution. The fitting function is
used to estimate the expected conflict time [36], which is shown in Equation (4).

dep 2
t arr −t −40.15
−( i i )
Tcon f lict = 52.4e 51.91 (4)

dep
where tiarr is the arrival time of flight i, and ti is the departure time of flight i.
The objective function (5) indicates the minimization of the conflict time (i.e., maxi-
mization of robustness):

Z3 = min ∑ ∑ Xik × Xjk × Yij × Tcon f lict (5)

i,j∈ F k ∈ G

where Yij is the decision variable, which is equal to one if flight i is consecutively followed
by flight j in the same gate and zero otherwise.
Int. J. Environ. Res. Public Health 2023, 20, 3952 5 of 12

The multiobjective gate assignment model is established to maximize the proportion

of flights assigned to the contact gate, minimize the aircraft taxiing fuel consumption, and
minimize the gate assignment robustness, which is shown in Equations (1), (3) and (5). The
constraint sets are as follows:
∑ xik = 1, ∀i, j ∈ F (6)
k∈G
dep dep
Xik X jk (ti − t arr
j )( t j − tiarr ) ≤ 0, ∀i, j ∈ F; k ∈ G (7)
dep
t arr


j − ti + 1 − Yij M ≥ T, ∀i, j ∈ F (8)

pi ≤ pk + (1 − Xik ) M, ∀i ∈ F; k ∈ G (9)
Xik , Yij ∈ {0, 1} (10)
where pi is the aircraft size; pk is the gate size; M is a large value; and T is the minimum
time interval.
Equation (6) indicates that each flight must and can only be assigned to one gate;
Equation (7) indicates that two flights cannot be assigned at the same gate at the same time;
Equation (8) indicates that two flights on the same gate shall meet a certain safety time
interval; and Equation (9) indicates that the gate type matches the aircraft type. Equation
(10) indicates that Xik , Yij are 0–1 variables.

3.2. Algorithm
Non-dominated sorting genetic algorithm II (NSGA-II) is deployed to solve the multi-
objective gate assignment model [35]. It refers to the process that classifies the population
by nondominated sorting of the population, calculates the crowding distance of popula-
tion individuals to maintain the diversity of the population, and obtains an approximate
solution when the termination condition is reached. The basic process framework of the
NSGA-II algorithm is shown in Figure 1.
The process of the NSGA-II is as follows:
Step 1: Code. The algorithm adopts decimal integer coding;
Step 2: Initialization. The algorithm initializes the population and generates the
first-generation subgroup through nondominated sorting, choosing, crossing, and chang-
ing meaning;
Step 3: Generation of a new population. The algorithm merges the parent generation
and the child generation to generate a new population;
Step 4: Determination of whether the termination conditions are met. If the end
conditions are met, stop the program, and otherwise, repeat Step 2 to Step 3.
The algorithm adopts the decimal integer encoding method. The genes on chromo-
somes represent gate numbers, and the chromosome length is the total number of flights.
An initial population with the scale of N is randomly generated as the first-generation
parent population, according to the constraints used for the gate, to the extent feasible.
The progeny population is then obtained by three basic operations: selection, crossover,
and mutation. By starting from the second generation, the parent population and the
progeny population are merged for fast nondominated sorting. At the same time, the
crowding degree of individuals in each nondominated layer is calculated. According to the
nondominated relationship and crowding degree of individuals, appropriate individuals
are selected to form a new population. This process continues until the conditions for the
end of the program are met. The composition of the new population of the algorithm is
shown in Figure 2, where the dominance relationship refers to the fact that each objective
function index of individual A is superior to that of another individual B, and thus A is
dominant to B; otherwise, it is a nondominated relationship, which is divided into several
layers according to the relationship between individuals. The individuals in the layer
are nondominated, and the individuals in the upper layer are dominated by those in the
lower layer.
Int.
[Link].
[Link].
[Link]
PublicHealth 2023,20,
Health2023, 20,3952
x FOR PEER REVIEW 66 of
of1213

Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 7 of 13

to B; otherwise, it is a nondominated relationship, which is divided into several layers

according to the relationship between individuals. The individuals in the layer are non-
dominated, and the individuals in the upper layer are dominated by those in the lower
layer.
[Link]
Figure Thebasic
basicprocess
processframework
frameworkofofthe
theNSGA-II
NSGA-IIalgorithm.
algorithm.

The process of the NSGA-II is as follows:

Step 1: Code. The algorithm adopts decimal integer coding;
Step 2: Initialization. The algorithm initializes the population and generates the first-
generation subgroup through nondominated sorting, choosing, crossing, and changing
meaning;
Step 3: Generation of a new population. The algorithm merges the parent generation
and the child generation to generate a new population;
Step 4: Determination of whether the termination conditions are met. If the end con-
ditions are met, stop the program, and otherwise, repeat Step 2 to Step 3.
The algorithm adopts the decimal integer encoding method. The genes on chromo-
somes represent gate numbers, and the chromosome length is the total number of flights.
An initial population with the scale of N is randomly generated as the first-generation
parent2. population,
Figure Example of a according
new speciesto the constraints
composition used for the gate, to the extent feasible.
of the NSGA-II.
The progeny population is then obtained by three basic operations: selection, crossover,
4. Experimental Results
[Link]
Experimental
mutation. By Results
starting from the second generation, the parent population and the prog-
4.1. Example Introduction
eny population are
4.1. Example Introduction merged for fast nondominated sorting. At the same time, the crowding
In this paper, Nanjing
degree of individuals in each Lukou International Airport
nondominated layer is(ZSNJ) was selected
calculated. to validate
According the
to the non-
model. In this
A paper,
total of 16Nanjing
airport Lukou
gates atInternational
the airport, Airport (ZSNJ)
including 13 was gates
contact selected to validate
(212–224) and
dominated relationship and crowding degree of individuals, appropriate individuals are
3the model.
remote
selected
A total
gates
to form(266,of267,
a new
16 and
airport gates
269),
population. wereat the airport,
Thisselected
processto
including
simulate
continues the13
until
contact
model. Thegates (212–224)
configuration
the conditions for the end
and
ofofthe3 remote gates (266, 267, and 269), were selected to simulate the model. The configu-
theairport
program is shown
are [Link] 3. The areas
composition marked
of the HS on theofmap
new population are conflict-prone
the algorithm is shown
rationofofthe
areas the airportInformation
is shown inon Figure
each 3. Theincluding
areas marked HSnumber,
on the map
type,are
andconflict-
in Figure 2,airport.
where the dominance gate,
relationship the fact
refers to the gate that each objective taxiing
function
prone areas
distance, is of the airport.
described in Information
Table 1. on each gate, including the gate number, type, and
index of individual A is superior to that of another individual B, and thus A is dominant
taxiing distance, is described in Table 1.
 4.1. Example Introduction
In this paper, Nanjing Lukou International Airport (ZSNJ) was selected to validate
the model. A total of 16 airport gates at the airport, including 13 contact gates (212–224)
and 3 remote gates (266, 267, and 269), were selected to simulate the model. The configu-
Int. J. Environ. Res. Public Health 2023, ration
20, 3952of the airport is shown in Figure 3. The areas marked HS on the map are conflict-7 of 12
prone areas of the airport. Information on each gate, including the gate number, type, and
taxiing distance, is described in Table 1.

Figure 3.
Figure 3. Gate
Gate location
location distribution
distributionofofthe
theairport
airportterminals at at
terminals ZSNJ.
ZSNJ.

[Link]
Table Information on the
the gates
gatesof
ofthe
theZSNJ
ZSNJairport.
airport.

Gate Number Number

Gate Number
Number Gate Type
Gate Type
Gate Size Arrival
Gate Size
Taxiing Distance (m) Departure
Arrival Taxiing Distance (m)
Taxiing Distance (m)
Departure Taxiing Distance (m)
212 G1 Near 3 1380 2450
212 G1 Near 3 1380 2450
213 G2 Near 3 1440 2390
213 G2 Near 3 1440 2390
214214 G3
G3 Near
Near 33 1100
1100 23302330
215215 G4
G4 Near
Near 33 1000
1000 22752275
216216 G5
G5 Near
Near 33 920920 22752275
217217 G6
G6 Near
Near 22 860860 18701870
218218 G7
G7 Near
Near 22 920920 18101810
219 G8 Near 2 980 1780
219 G8 Near 2 980 1780
220 G9 Near 2 1030 1730
221220 G9
G10 Near
Near 22 1030
1100 17301670
222 G11 Near 2 1150 1580
223 G12 Near 2 1000 1150
224 G13 Near 2 1050 1320
266 G14 Far 3 1980 2880
267 G15 Far 3 1900 2950
269 G16 Far 3 2100 3000
Note: In the gate size column, 1 represents a small gate, 2 represents a medium gate, and 3 represents a large gate.

The flight operation data of the airport were also extracted. A total of 136 domestic
flights taking off and landing at the airport from 8:30 to 18:00 on 1 September 2019, namely
68 domestic arrival and departure flights, were selected for simulation verification. The
specific flight information is shown in Table 2. According to the aviation engine emission
database released by the International Civil Aviation Organization (ICAO), the fuel flow
rate of the engine in the idle state corresponding to each flight type can be obtained, as
shown in Table 3.

Table 2. Example of arrival and departure flight data at the ZSNJ airport.

Flight Pair Arrival Flight Arrival Departure Flight Departure Engine Fuel Flow
Aircraft Size Engine
Number Number Time Number Time Rate (kg/s)
1 ZH9855 8:35 ZH9856 9:35 2 2 0.109
9 CZ6189 9:15 CZ6189 10:20 2 2 0.138
18 3U8923 10:30 3U8924 11:45 3 2 0.228
21 ZH9843 11:10 ZH9844 12:25 3 2 0.270
52 CZ6971 16:25 CZ6972 17:45 2 2 0.109
Note: In the Aircraft size column, 1 represents a small aircraft, 2 represents a medium aircraft, and 3 represents a
large aircraft.
Int. J. Environ. Res. Public Health 2023, 20, 3952 8 of 12

Table 3. Engine fuel flow rates under idle conditions.

Aircraft Type Engine Engine Type Engine Fuel Flow Rate (kg/s)
32 N 2 LEAP-1A26CJ 0.091
A319 2 CFM56-5B5/P 0.102
A320 2 CFM56-5A3 0.104
A321 2 V2530-A5 0.138
A322 2 CFM56-5B4/P 0.102
A332 2 CF6-80C2B4 0.228
A333 2 Trent 772B-60 0.270
B737 2 CFM56-3C-1 0.124
B738 2 CFM56-7B24/2 0.109
B789 2 GEnx-1B70/72/P2 0.201

4.2. Results and Discussion

The average taxiing speed of the aircraft was set as 15 km/h, and the shortest time
interval of the same position was set as 15 min [6]. The maximum number of iterations was
200. This paper used the Visual Studio 2013 platform to achieve the NSGA-II algorithm.
A total of 17 groups of Pareto front solutions were obtained, and their distribution
Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 9 of 13
and fitting surfaces are shown in Figure 4. The Pareto front solutions are defined as non-
dominated solutions and have the least amount of model goal conflict, which provides
decision-makers with a better choice space. It can be seen in Figure 4 that a more robust
gate
gate assignment plan planwill
willrequire
requiremore
moreidleidle gate
gate usage,
usage, which
which results
results in more
in more aircraft
aircraft being
being
assignedassigned to remote
to remote gates. gates.
It will It will decrease
decrease the proportion
the proportion ofassigned
of flights flights assigned to gates
to contact con-
tact gates and increase the total fuel consumption and carbon emissions. In
and increase the total fuel consumption and carbon emissions. In the resource-constrained the resource-
constrained
situation, the situation,
proportionthe of
proportion of flights
flights assigned to assigned to contact
contact gates gates
and gate and gate assign-
assignment robust-
ment robustness
ness present present arelationship.
a negative negative relationship.
Therefore,Therefore, gate assignment
gate assignment needs to needs
improve to im-
the
prove the utilization
utilization rate of therate of thegate
contact contact
and gate
reduceandthe
reduce thecarbon
overall overallemissions
carbon emissions while
while keeping
keeping the assignment
the assignment robustness
robustness within awithin a reasonable
reasonable range. range.

Figure 4.
Figure Paretofront
4. Pareto front solution.
solution.

The three best solutions were selected to be compared with the original scheme (OS)
The three best solutions were selected to be compared with the original scheme (OS)
in Table 4. The OS of gate assignment is the original gate assignment provided by the
in Table 4. The OS of gate assignment is the original gate assignment provided by the
airport. As seen in Table 4, the three assignment solutions do not dominate each other.
airport. As seen in Table 4, the three assignment solutions do not dominate each other.
Compared with the OS, the results of the gate assignment model have been improved for
Compared with the OS, the results of the gate assignment model have been improved for
different objectives.
different objectives.
The first one is the result when Z1 and Z2 reach the optimum at the same time. The total
The first one is the result when Z1 and Z2 reach the optimum at the same time. The
taxiing fuel consumption and carbon emissions are significantly reduced compared with
total taxiing fuel consumption and carbon emissions are significantly reduced compared
the actual assignment scheme, but the estimated total conflict time is 515 min, resulting in a
with the actual assignment scheme, but the estimated total conflict time is 515 min, result-
ing in a high cost of flight delay and additional carbon emissions. Then, the result of opti-
mal Z3 is presented. It has the lowest robust cost, but the remote gate rate and fuel con-
sumption cost are too high. The result of optimizing the three objectives (Z1 + Z2 + Z3) is
shown in Table 4. Compared with the original scheme, it increased the rate of flight as-
 Int. J. Environ. Res. Public Health 2023, 20, 3952 9 of 12

high cost of flight delay and additional carbon emissions. Then, the result of optimal Z3 is
presented. It has the lowest robust cost, but the remote gate rate and fuel consumption cost
are too high. The result of optimizing the three objectives (Z1 + Z2 + Z3 ) is shown in Table 4.
Compared with the original scheme, it increased the rate of flight assigned to the contact
gate by 7.36% and reduced the total taxiing fuel consumption by 550.08 kg. Moreover, it
reduced the total taxiing carbon emissions and fuel consumption by 2552.94 kg and 8.09 kg,
respectively. However, it increased the estimated total conflict time by 50 min. In summary,
the proposed gate assignment model can reduce carbon emissions to different degrees.

Table 4. Pareto front solutions based on the NSGA-II algorithm.

Average Taxiing Fuel Total Taxiing Carbon

Objective(s) Z1 Z2 Z3
Consumption (kg) Emissions (kg)
OS 85.29% 12,401.44 284.42 182.37 57,555.16
Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 10 of 13
Z1 + Z2 94.12% 11,670.32 515.66 171.62 54,162.01
Z3 83.82% 12,534.78 221.68 184.34 58,173.98
Z1 + Z2 + Z3 92.65% 11,851.36 334.57 174.28 55,002.22

Figure 6. As shown in Figures 5 and 6, only five flight pairs with a long overpass time are
A Gantt diagram of the original gate assignment is shown in Figure 5. The Gantt
assigned to the remote gates, which significantly improves the utilization efficiency of the
diagram of gate assignment with optimization of three objectives (Z1 + Z2 + Z3 ) is shown in
contact
Figuregates.
6. As Since
shownthe model 5established
in Figures and 6, only infivethis paper
flight pairsfully
withconsiders the factors
a long overpass time areof taxi-
ingassigned
distancetoand the aircraft type,which
remote gates, the mean and variance
significantly improves of fuel
the consumption are significantly
utilization efficiency of the
reduced
contact gates. Since the model established in this paper fully considers the factorsthat
compared with the original assignment scheme, which indicates the taxiing
of taxiing
fuel consumption
distance of each
and aircraft type,flight is relatively
the mean balanced.
and variance of fuelIf consumption
we take Flight are18 as an example,
significantly
thereduced
aircraftcompared
type is a with the original assignment
medium-length wide-bodyscheme, aircraft,which indicates that
an A330-200. the taxiing
Its engine fuel flow
fuel consumption of each flight is relatively balanced. If we take Flight
rate is significantly higher than that of other small and medium aircraft, which greatly 18 as an example,
the aircraft
impacts type is
the total a medium-length
carbon [Link]-body
After the aircraft, an A330-200.
optimization in thisItspaper,
enginethefuelflight
flow was
rate is significantly higher than that of other small and medium aircraft, which greatly
changed from remote gate 14 to contact gate 4, and the fuel consumption was reduced by
impacts the total carbon emissions. After the optimization in this paper, the flight was
205.42 kg. The
changed fromfuel-saving
remote gateeffect
14 toof a single
contact gateflight
4, and wasthesignificant, and thewas
fuel consumption overall fuel con-
reduced
sumption
by 205.42waskg. effectively reduced.
The fuel-saving effectThus, carbon
of a single flightemissions are significantly
was significant, reduced.
and the overall fuel
consumption was effectively reduced. Thus, carbon emissions are significantly reduced.

Figure 5. Gantt chart of the original gate assignment.

Figure 5. Gantt chart of the original gate assignment.
 Figure
Int. J. Environ. Res. Public Health 2023, 5. Gantt
20, 3952 chart of the original gate assignment. 10 of 12

Figure
Figure 6. 6. Ganttchart
Gantt chartof
of the
the NSGA-II
NSGA-II algorithm
algorithmgate assignment
gate solution.
assignment solution.
In summary, the simulation optimization result indicates that the proposed model can
effectively reduce carbon emissions and simultaneously improve the utilization efficiency
of contact gates, airport operation efficiency and robustness, thereby enhancing passenger
satisfaction and reducing delay costs. In addition, the results show that the NSGA-II
algorithm can solve multiobjective models efficiently by fast non-dominated sorting and
crowding sorting processes.

5. Conclusions
This paper proposes a multiobjective gate assignment model that considers carbon
emissions to facilitate environmentally friendly airport construction. The proportion of
flights assigned to the contact gate, aircraft taxiing fuel consumption, and gate assignment
robustness are considered in the model to reduce carbon emissions. Multiple objectives
of the gate assignment model interact with each other, which makes it challenging to
explore the optimal solution. Therefore, in this research, the NSGA-II was used to solve
the problem. Numerical studies have shown that the proposed gate assignment model
could achieve better performance in decreasing carbon emissions compared with the
original gate assignment plan. In addition, the proposed model can effectively reduce
the walking distance of passengers, the operation cost of airlines, and the conflict time at
the airport surface to different degrees. It has good market prospects for improving the
economic benefits of airports and airlines while improving the utilization rate of airport
gate resources.
Although the preliminary results are promising, there is still much research to be
done. In this paper, only the partial process of gate assignment at the airport surface
is considered, and the taxi route of the aircraft is assumed to be fixed. In addition, the
mathematical models established in the current paper only considered deterministic cases,
while several stochastic events, such as weather conditions, flight delays, and the impact of
ground support operations on the apron, were not fully considered. Future research will
focus on the joint scheduling of runway, taxiway and gate operations, considering multiple
uncertainties to fully achieve the goals of energy conservation and emission reduction in
airport operation.
Int. J. Environ. Res. Public Health 2023, 20, 3952 11 of 12

Author Contributions: This work is the result of collaboration among Y.S. (Yixuan Shan), Y.S. (Yuwei
Shao), Q.Y. and Y.J. All authors have equally contributed, reviewed, and improved the manuscript.
All authors have revised the final manuscript. All authors have read and agreed to the published
version of the manuscript.
Funding: This research was funded by Undergraduate Innovation Program (2022CX007048).
Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: No new data were created or analyzed in this study. Data sharing is
not applicable to this article.
Acknowledgments: Thanks to reviewers for their constructive comments on research design
and writing.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Federal Aviation Administration Office of Environment and Energy. Aviation Emissions and Air Quality Handbook Version 3 Update
1; FAA: Washington, DC, USA, 2016.
2. Long, J.L. Strategic thinking on developing new energy of air transport under low carbon economy. China Transp. Rev. 2013, 4, 9–13.
3. Civil Aviation Administration of China. Outline of Action for the Construction of Four Types of Airports in China Civil Aviation; CAA:
Beijing, China, 2020.
4. Winther, M.; Kousgaard, U.; Ellermann, T.; Massling, A.; Ketzel, M. Emissions of NO x, particle mass and particle numbers from
aircraft main engines, APU’s and handling equipment at Copenhagen Airport. Atmos. Environ. 2015, 100, 218–229. [CrossRef]
5. Sznajderman, L.; RamírezDíaz, G.; Bernardi, C.A.D. Influence of the Apron Parking Stand Management Policy on Aircraft and
Ground Support Equipment (GSE) Gaseous Emissions at Airports. Aerospace 2021, 8, 87. [CrossRef]
6. Xue, Q.W.; Jiang, Y.; Lui, Z.L.; Zeng, L.Y. Gate Assignment Based on Airline Operation Cost and Fairness. Aeronaut. Comput.
Technol. 2016, 46, 64–69.
7. Shen, X.Y.; Yu, H.W. Research on multi-objective gate assignment based on aircraft matching degree. World Transp. Congr.
2018, 47, 445–457.
8. Jiang, Y.; Hu, Z.T.; Tong, C.; Liu, Z.Y.; Chen, L.L.; Zhang, H.H. An Optimization Model for Gate Re-assignment under Flight
Delays. J. Transp. Syst. Eng. Inf. Technol. 2020, 20, 185–190, 217.
9. Yuan, Y.; Zhai, H.X.; Sun, C.L. Gate Assignment of Freight Flight Considering Efficiency Operation Support. J. Nanjing Univ.
Aeronaut. Astronaut. 2021, 53, 934–940.
10. Yan, P.; Yuan, Y. Joint Optimization of Flight Taxiing Path Planning and Gate Allocation. Control Eng. China 2021, 28, 464–470.
11. Xiong, J.; Zhang, C. Airport Gate Assignment with Airplane Taxing Cost Analysis. J. Transp. Syst. Eng. Inf. Technol.
2010, 10, 165–170.
12. Wang, X.T.; Tian, Y.; Wan, L.L.; Yang, Y. Research on Robustness of Gate Assignment Based on Column Generation Methods.
J. Wuhan Univ. Technol. Transp. Sci. Eng. 2015, 39, 171–174, 179.
13. Li, Q.W. Research on Optimization Models for Airport Gate Assignment Problem; Beijing Jiaotong University: Beijing, China, 2017.
14. Daş, G.S. New Multi-Objective Models for the Gate Assignment Problem. Comput. Ind. Eng. 2017, 109, 347–356. [CrossRef]
15. Genc, H.M.; Erol, O.; Eksin, I.; Berber, M.; Gueleryuez, B.O. A stochastic neighborhood search approach for airport gate assignment
problem. Expert Syst. Appl. 2012, 39, 316–327. [CrossRef]
16. Kumar, V.P.; Bierlaire, M. Multi-objective airport gate assignment problem in planning and operations. J. Adv. Transp. 2014, 48, 902–926.
[CrossRef]
17. Zheng, W.J.; Le, M.L.; Shao, J.J.; Li, T. Research on Assignment of Gate Assignment Based on Resource Utilization Efficiency.
Aviat. Comput. Technol. 2019, 49, 70–73, 78.
18. Marina, K.; Laura, L. Biofuels in aviation: Fuel demand and CO2 emissions evolution in Europe toward 2030. Transp. Res. Part D
Transp. Environ. 2016, 46, 166–181.
19. Wan, L. Research on the Optimization Method of Air Traffic Operation for Environmental Protection; Nanjing University of Aeronautics
and Astronautics: Nanjing, China, 2015.
20. Kuzu, S.L. Estimation and dispersion modeling of landing and take-off (LTO) cycle emissions from Atatürk International Airport.
Air Qual. Atmos. Health 2017, 11, 153–161. [CrossRef]
21. Gao, W.; Cui, Y.; Kang, D. Optimization of airport ground operation for energy conservation and emission reduction.
J. Transp. Inf. Saf. 2019, 37, 128–135.
22. Evertse, C.; Visser, H.G. Real-time airport surface movement planning: Minimizing aircraft emissions. Trans. Res. Part C Emerg. Technol.
2017, 79, 224–241. [CrossRef]
23. Soltani, M.; Ahmadi, S.; Akgunduz, A.; Bhuiyan, N. An eco-friendly aircraft taxiing approach with collision and conflict avoidance.
Transp. Res. Part C Emerg. Technol. 2020, 121, 102872. [CrossRef]
Int. J. Environ. Res. Public Health 2023, 20, 3952 12 of 12

24. Balakrishnan, H. Queuing Models of Airport Departure Processes for Emissions Reduction; American Institute of Aeronautics &
Astronautics Guidance, Navigation, and Control Conference: Chicago, IL, USA, 2009.
25. Li, N.; Zhang, H. Calculating aircraft pollutant emissions during taxiing at the airport. Acta Sci. Circumstantiae 2017, 37, 1872–1876.
26. Zhang, H. Research on the Strategy of Aircraft Low Emission Taxiing on the Airports; Civil Aviation University of China: Tianjing,
China, 2017.
27. Lu, M.Y. Research on Jointed Scheduling Optimization of Busy Airport Taxiway and Gate; Civil Aviation Flight University of China:
Tianjing, China, 2022.
28. Tang, X.W.; Chen, Z.; Zhang, S.R.; Ye, D. Impact of Apron Spatial Configuration on Flight Departure Taxi Time at Busy Airports.
J. Transp. Syst. Eng. Inf. Technol. 2022, 22, 309–317.
29. Li, R.N.; Feng, X. Optimization of Flight Area Layout with Minimum Fuel Consumption Based on Improved Genetic Algorithm.
J. Chongqing Jiaotong Univ. Nat. Sci. 2022, 10, 1–8.
30. Gök, Y.S.; Padrón, S.; Tomasella, M.; Guimarans, D.; Ozturk, C. Constraint-based robust planning and scheduling of airport apron
operations through simheuristics. Ann. Oper. Res. 2022, 320, 795–830. [CrossRef]
31. Yin, S.; Han, K.; Ochieng, W.Y.; Sanchez, D.R. Joint apron-runway assignment for airport surface operations. Transp. Res. Part B
2022, 156, 76–100. [CrossRef]
32. Wen, J.; Sun, H.; Xu, J.; Liang, Z.J. Study of the Gate Assignment in Airport Based on Fixed Job Scheduling Algorithm. Syst. Eng.
2004, 7, 102–106.
33. Liu, J.L. Research on the Airport Gate Assignment Based on Difference Evolution Algorithm; Civil Aviation Flight University of China:
Tianjing, China, 2022.
34. Liu, J.L.; Wang, J.H.; Liang, H.J. Research on Parking Space Allocation Based on Genetic Algorithm. Comput. Inf. Technol.
2020, 28, 14–15, 56.
35. Xu, S.M.; Jiang, Y.; Wang, H.; Chen, L.L. Multi-objective Gate Assignment Modeling and Simulation Based on NSGA-II.
J. Nanjing Univ. Aeronaut. Astronaut. 2018, 50, 823–828.
36. Liu, Z. Research on Optimization of Airport Gate Assignment Based on Column Generation; Nanjing University of Aeronautics and
Astronautics: Nanjing, China, 2022.

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.