sustainability

Article
Design of An Integration Model for Air Cargo
Transportation Network Design and Flight
Route Selection
Carman K.M. Lee 1 , Shuzhu Zhang 2, * and Kam K.H. Ng 3

1 Department of Industrial Engineering, the Hong Kong Polytechnic University, Hung Hom, Kowloon,
Hong Kong 999077, China; [email protected]
2 Department of Information Management and Engineering, Zhejiang University of Finance & Economics,
Hangzhou 310018, China
3 Department of Industrial Engineering, the Hong Kong Polytechnic University, Hung Hom, Kowloon,
Hong Kong 999077, China; [email protected]
* Correspondence: [email protected]




Received: 20 August 2019; Accepted: 19 September 2019; Published: 23 September 2019


Abstract: Air cargo transportation is an essential component in the freight transportation market,
primarily due to the transportation requirements of time-sensitive products. Air cargo transportation
plays an increasingly important role alongside economic development. Cargo flight network
design and fleet routing selection significantly affect the performance of the air cargo transportation.
In this research, we propose an integrated model simultaneously considering cargo flight network
design and the fleet routing selection for the air cargo transportation. Two transportation
modes, the direct transportation mode in point-to-point networks and the transshipment mode
in hub-and-spoke networks, are compared. In order to solve the proposed optimization problem,
a swarm-intelligence-based algorithm is adapted. Numerical experiments were conducted to examine
and validate the effectiveness and efficiency of the proposed model and algorithm. The computational
results suggest that the proper settings of hub and transshipment route selection in an air cargo
transportation network can significantly reduce the transportation cost, which can provide practical
managerial insights for the air cargo transportation industry.

Keywords: air cargo transportation; network design; fleet route selection; hub-and-spoke network

1. Introduction
The importance of air cargo transportation rises in step with economic development. Air cargo
is an essential part of the global trading system [1]. The IATA annual review shows that in 2016,
airlines transported 53.9 million metric tons of goods, representing more than 35% of global trade by
value but less than 1% of world trade by volume, and it is reported that the average yearly growth
of air cargo in 2017 and 2018 were 9.7% and 3.4%, respectively, among which the latter is partially
due to protectionist policies [2]. Moreover, according to the world air cargo forecast provided by
Boeing, the world air cargo traffic will more than double over the next twenty years. Among all the
transported air cargoes, a large proportion is accomplished using mixed cargo–passenger aircrafts,
especially in developing countries, which results in the dominance of air passenger transportation
management over air cargo transportation [3]. The application of mixed cargo–passenger aircrafts
favors the requirements of passengers as the first priority, and cannot satisfy the rapid development
of air cargo transportation [4,5]. Therefore, the utilization of dedicated cargo aircraft for air cargo
transportation is gradually becoming popular due to the development of e-commerce and express
delivery. For example, Shunfeng (one major logistics company in China) has not only purchased more

Sustainability 2019, 11, 5197; doi:10.3390/su11195197 www.mdpi.com/journal/sustainability

Sustainability 2019, 11, 5197 2 of 12

than 40 airplanes dedicated to air cargo transportation, but also built its own air cargo airport in Hubei
province acting as its international logistics hub. The development of air cargo transportation is an
inevitable trend along with regional and national economic development [6].
Air cargo transportation differs from air passenger transportation in many respects. Kasilingam [7]
firstly explored similarities and differences between passenger and cargo capacity management in
the airline context. The hub selection of passenger service networks usually involves long-term
planning; in contrast, cargo movement is more flexible and dynamic [8]. Moreover, passengers are
more time-sensitive than cargo, which means that passengers prefer direct flights, while cargo may
be transshipped multiple times as long as it can be delivered on time. Therefore, existing passenger
transportation networks and flight routes need to be adjusted in order to meet the latest requirements
of air cargo transportation.
Airline network design underpins the performance of air cargo transportation. Two types
of airline network (i.e., point-to-point and hub-and-spoke) and their combination are commonly
observed in practice [9,10]. The comparison of these two network types has been conducted in many
studies. For example, Lin et al. [11] compared the economic effect of center-to-center directs on
hub-and-spoke networks for air express common carriers. Furthermore, they proposed an integral
constrained generalized hub-and-spoke network design model in a capacitated and directed network
configuration [12]. The point-to-point network is also known as direct transportation, which is
suitable for the transportation of time-sensitive goods. Comparatively, the hub-and-spoke network
suits transshipment transportation as it is capable of flow consolidation and dissemination [13]. The
construction of the hub-and-spoke network comprises two components: the determination of hubs and
the allocation of non-hub nodes to hubs. Oktal and Ozger [14] conducted a case study to determine the
location of hub airports in air cargo transportation. O’Kelly et al. [15] firstly compared the hub network
design with single and multiple allocation strategies, then further analyzed the flow economies of scale
under different location strategies [16]. Bryan and O’Kelly [17] provided an analytical review regarding
the design of hub-and-spoke networks. Kim [18] once introduced a p-hub protection model aiming to
maximize the total interacting traffic over a set of origin–destination nodes based on different routing
strategies. In addition to the air cargo transportation network design, the selection of an air cargo
transportation route can be significantly different from air passenger transport routes. Kasilingam [7]
pointed out that cargo can be shipped among different routes as long as it arrives at its destination on
time. Especially in hub-and-spoke networks, the transshipment transportation of air cargo becomes
a promising and popular trend in view of the low transportation cost. Ohashi et al. [19] applied an
aggregate multinomial logit model to identify the critical factors influencing air cargo transshipment
route choice decisions. Kim and Park [20] investigated the degree and quality of transshipment
connectivity in an air cargo hub.
In this research, we propose an integrated model for air cargo transportation network design and
flight route selection in view of the practical requirements of air cargo transportation. The balance
between time-sensitivity and flow economy of scale can be handled by the integration of point-to-point
network and hub-and-spoke network, and the corresponding flight route selection and transshipment
arrangements. Due to the NP complexity feature of the proposed model, a swarm-intelligence-based
approach is employed in this research. Swarm-intelligence-based approaches derive from the collective
and intelligent foraging or hunting behavior of social insects and group animals. Among multiple
swarm-intelligence-based approaches, the artificial bee colony (ABC) algorithm is one popular example,
which is capable of balancing the diversified and intensified search performance by the proper
adjustment of the involved parameters and control mechanisms, and can reach convergence in an
acceptable computational time. A succinct solution representation scheme containing adequate
information on hub determination and route selection was designed using the ABC algorithm’s
search process. After that, a number of numerical experiments were conducted to examine and
validate the performance of the proposed mathematical model and algorithm. The results indicate that
Sustainability 2019, 11, 5197 3 of 12

the application of a hub-and-spoke network and transshipment can generate substantial savings in
transportation costs, where the number of hubs and their determination are the critical factors.
The remainder of this paper is organized as follows. After the introduction of air cargo
transportation and a brief review of the relevant literature, a mathematical formulation of the
emerging problem is provided in Section 2. Section 3 describes the solution approach. Numerical
experiments were conducted as presented in Section 4 to validate the effectiveness and efficiency of the
proposed model and method. Finally, the conclusions are drawn in Section 5.

2. Problem Formulation
In this research, in order to construct the air cargo transportation network, full network connectivity
was initially assumed. Regarding the hub allocation, we assumed that one non-hub node was associated
with at most one designated hub, and that one hub node served the transshipment requirements
of multiple non-hub airports within its coverage. The air cargo transportation in between each
origin–destination pair had three possible scenarios, that is, direct transportation as yij , one-time
transshipment at a hub as yik j , and two transshipments at two consecutive hubs as yiklj , respectively.
The shipment requirement from a specific origin node to a particular destination node cannot be split.
A homogenous fleet of flights was assumed, with the same capacity and speed. Both the non-hub node
and hub had enough flights to handle the transportation and transshipment demand. The operational
time at the origin and destination nodes was not considered, but the transshipment time at the hub
was considered if transshipment was needed. The problem was to determine which airport(s) could be
redesigned as a hub airport, which non-hub airport(s) it/they served, and the transportation routes
between each origin–destination (OD) pair.
The following notations were used to construct the air cargo transportation network.
N The set of nodes
fk , ∀k ∈ N The cost of operating hub k
p The predetermined number of hubs
dmax The maximum distance between a non-hub node and its associated hub
dij , ∀i, j ∈ N The distance between nodes i and j
cij , ∀i, j ∈ N The transportation cost between nodes i and j per shipment unit, per distance unit
α The discount factor for inter-hub transportation due to the economies of scale (0 < α < 1)
The discount factor for the transportation cost between a non-hub node and a hub due to
β
the economies of scale (0 < β < 1)
ck , ∀k ∈ N The transshipment cost at hub k per shipment unit
wij , ∀i, j ∈ N The quantity of shipments from node i to node j

The decision variables used in this research are described as follows.

xk ∈ {0, 1}, ∀k ∈ N Whether node k is a hub or not
zik ∈ {0, 1}, ∀i, k ∈ N Whether the non-hub node i is associated with hub k or not
yij ∈ {0, 1}, ∀i, j ∈ N The direct transportation between the origin node i and the destination node j
The transportation between the origin node i and the destination node j,
yik j ∈ {0, 1}, ∀i, k, j ∈ N
transshipped at hub k
The transportation between the origin node i and the destination node j,
yiklj ∈ {0, 1}, ∀i, k, l, j ∈ N
transshipped at hubs k and l

The complete mathematical problem is described as follows.

Sustainability 2019, 11, 5197 4 of 12

Objective
P P P  
β cik dik + ck j dk j qi j yik j
P P
minz = ci j di j qi j yi j +
i∈N j∈N i∈N j∈N k∈N
P P P P
β·cik dik + α·ckl dkl + β·cl j dl j qi j yikl j +
P P P
+ ck qi j yik j (1)
i∈N j∈N k∈N l∈N k∈N i∈N j∈N
P P P P P
+ cl qi j yikl j + fk xk
l∈N i∈N j∈N k∈N k∈N
Subject to
zik ≤ xk , ∀i, k ∈ N (2)
P
zik = 1, ∀i ∈ N (3)
k∈N P
xk ≤ p (4)
k∈N
dik zik ≤ dmax , ∀i, k ∈ N (5)
yik j ≤ zik , ∀i, j, k ∈ N (6)
yik j ≤ z jk , ∀i, j, k ∈ N (7)
yik j ≤ zik + z jk − 1, ∀i, j, k ∈ N (8)
y ≤ xk , ∀i, j, k ∈ N (9)
P ik j
yikl j ≤ zik , ∀i, j, k ∈ N (10)
l∈N
P
yikl j ≤ z jl , ∀i, j, l ∈ N (11)
k∈N
yikl j ≤ zik + z jl − 1, ∀i, j, k, l ∈ N (12)
yikl j ≤ xk + xl − 1, ∀i, j, k, l ∈ N (13)
P P P
yi j + yik j + yikl j = 1, ∀i, j ∈ N (14)
   k∈N k∈N l∈N 
β cik dik + ck j dk j wi j yikl + β·cik dik + α·ckl dkl + β·cl j dl j wi j yikl j ≤ ci j di j wi j yi j , ∀i, j, k, l ∈ N (15)
   
β cik dik + ck j dk j wi j yikl ≤ β·cik dik + α·ckl dkl + β·cl j dl j wi j yikl j , ∀i, j, k, l ∈ N (16)
xk , zik , yi j , yik j , yikl j ∈ {0, 1}, ∀i, j, k, l ∈ N (17)

The objective function (1) is to minimize the entire cost, which consists of the air cargo transportation
cost as illustrated in the first three terms, the transshipment cost as illustrated in the fourth and fifth
terms, and the hub operating cost as illustrated in the sixth term. Constraint (2) indicates that a non-hub
node can only be associated with an open hub. Constraint (3) reveals that a non-hub node can only
be assigned to one hub. Constraint (4) means the number of hubs cannot exceed a predetermined
number. Constraint (5) indicates that the distance between a non-hub node and its associated hub
shall be within a reasonable distance range. Constraints (6) and (7) describe one-time transshipment
transportation provided that the origin and destination node are associated with the same hub node.
Constraints (8) and (9) describe the corresponding relationship between decision variables. Constraints
(10) and (11) describe the two-time transshipment transportation provided that non-hub nodes i and
j are associated with hub nodes k and l, respectively. Constraints (12) and (13) describe the relation
between decision variables in the case of two-time transshipment. Constraint (14) expresses that there
is only one option for the transportation between each OD pair, which can be direct transportation, or
one-time transshipment transportation, or two times transshipment transportation. Constraints (15)
and (16) indicate that direct transportation is the highest-cost option, and the transshipment operation
can lower the transportation cost. Constraint (17) indicates that all decision variables are binary.

3. Solution Approach
In order to solve the proposed model, it was necessary to design a proper code mechanism for
representing a candidate solution. A solution scheme with lengthy dimensions would cause a slow
convergent search performance. Therefore, the proper solution scheme should be coded as succinctly
as possible, yet in way sufficient to reconstruct the complete solution. After determination of the
solution representation scheme, the ABC algorithm was employed with appropriate combinations of
adjustment parameters and control mechanisms in order find the optimal solution.
Sustainability 2019, 11, 5197 5 of 12
Sustainability 2019, 11, x FOR PEER REVIEW 5 of 12

3.1. Solution
Solution Representation
Representation
The proposed
proposed integrated
integratedmodel
modelcomprises
comprises twotwo interrelated
interrelated components:
components: the the network
network design
design and
and the route
the route selection.
selection. The The
networknetwork
designdesign
can becandivided
be dividedinto into
two two separate
separate sections,
sections, the first
the first beingbeing
hub
hub determination,
determination, and and the second
the second the allocation
the allocation between
between hub hub
nodes nodes and non-hub
and non-hub nodes.
nodes. OnceOnce the
the hub
hub information
information is determined,
is determined, the transportation
the transportation cost can ifbe
cost can be calculated thecalculated if the product
product transportation goes
transportation
through transshipment.goes through transshipment.
In this research, weIn this research,
designed we designed
a solution scheme as a solution
S = {Sh , scheme as 𝑆 =S
Sr }, in which
{𝑆 , 𝑆 }, in
denotes thewhich
candidate𝑆 denotes
solution,theScandidate
h denotes the solution, 𝑆 denotes
hub determined in the
thishub determined
network, in this network,
and Sr denotes whether
each 𝑆ODdenotes
and whether
transportation each OD transportation
is transshipped or not. Figure is 1transshipped or not.representation
illustrates a solution Figure 1 illustrates
example. a
solution representation
In this example, example.
the network In this example,
is comprised of 10 nodesthe indexed
networkfrom is comprised of 10which
0 to 9, among nodesnodeindexed
1, 3,
from
7, and0 8toare9, among
designatedwhichasnode 1, 3, 7, and
hub nodes. 8 are designated
Concerning as hubbetween
the allocation nodes. Concerning
hub nodes the andallocation
non-hub
between hub nodes
nodes, a non-hub andisnon-hub
node allocatednodes, a non-hub
to its closest node is
hub node, allocated
which to itsfrom
is drawn closest
thehub node,hub
available which is
node
drawn from the available hub node list 𝑆 . For the transportation
list Sh . For the transportation mode, Sr contains the information of whether transshipment is used. mode, 𝑆 contains the
information
For example,ofinwhether this case,transshipment
nodes 0 and 1 is used. For
transport example,
products in this
to their case, nodes 0destination
corresponding and 1 transport
nodes
products to their corresponding
using transshipment, and node 2 destination
transports its nodes using
product transshipment,
to its corresponding and node 2 transports
destination its
using direct
product to its corresponding destination using direct transportation.
transportation. A similar mechanism was applied to other nodes to calculate their OD transportationA similar mechanism was
applied
distanceto andother
cost.nodes to calculate their OD transportation distance and cost.

𝑆ℎ = 1 3 7 8

𝑆𝑟 = 1 1 0 1 0 1 1 0 0 0

0 1 2 3 4 5 6 7 8 9

Figure 1.
Figure Solution illustration
1. Solution illustration example.
example.

3.2. ABC Algorithm

3.2. ABC Algorithm
The ABC algorithm was inspired by the searching behavior of honey bees, which can find the
The ABC algorithm was inspired by the searching behavior of honey bees, which can find the
most promising food source in a large neighborhood area [21]. A bee colony is composed of three
most promising food source in a large neighborhood area [21]. A bee colony is composed of three
types of bees: scout bees, employed bees, and onlooker bees. Different types of bees carry out different
types of bees: scout bees, employed bees, and onlooker bees. Different types of bees carry out
functions. The implementation of the ABC algorithm comprises four distinct phases. Phase 1 is the
different functions. The implementation of the ABC algorithm comprises four distinct phases. Phase
initialization phase. In this phase, a number of scout bees are sent out to search for food sources
1 is the initialization phase. In this phase, a number of scout bees are sent out to search for food
in a certain area. One food source corresponds to one solution, and the richness of the food source
sources in a certain area. One food source corresponds to one solution, and the richness of the food
corresponds to the fitness of the solution. Phase 2 is the employed bee phase. Employed bees are
source corresponds to the fitness of the solution. Phase 2 is the employed bee phase. Employed bees
designed to exploit the found food sources. One employed bee exploits one food source, and during
are designed to exploit the found food sources. One employed bee exploits one food source, and
the exploitation, it also searches the neighborhood of the current food source to determine whether a
during the exploitation, it also searches the neighborhood of the current food source to determine
better food source exists. If a better food source is found, the corresponding employed bee replaces the
whether a better food source exists. If a better food source is found, the corresponding employed bee
current food source with the better one. Phase 3 is the onlooker bee phase. The onlooker bees receive
replaces the current food source with the better one. Phase 3 is the onlooker bee phase. The onlooker
food source information from the employed bees, and then decide whether to follow certain employed
bees receive food source information from the employed bees, and then decide whether to follow
bees and repeat the exploitation and exploration operations. Phase 4 is the scout bee phase. As the
certain employed bees and repeat the exploitation and exploration operations. Phase 4 is the scout
food sources are gradually consumed, they are abandoned once they are depleted. In this case, scout
bee phase. As the food sources are gradually consumed, they are abandoned once they are depleted.
bees are sent out to find alternative food sources. The pseudo code of the ABC algorithm is presented
In this case, scout bees are sent out to find alternative food sources. The pseudo code of the ABC
in Figure 2.
algorithm is presented in Figure 2.
The proper balance between diversification and intensification of the search process is critical to
find promising solutions [22]. In the ABC algorithm, this balance is accomplished by the algorithm’s
parameter settings and control mechanism. The first parameter is the solution number (SN). More
initial solutions indicate a more diversified search. The second parameter is the abandonment criterion,
denoted as the limit. A higher abandonment criterion means that the algorithm will preserve the
current solution by continuing to dig into its neighborhood search area for better solutions. In this case,
the algorithm performs as a more intensified search. Moreover, both the greedy selection mechanism
in the neighborhood search process and the probabilistic selection of solutions in the onlooker bee
Sustainability 2019, 11, 5197 6 of 12

phase can aggravate the intensified search. The search performance using different combinations of
parameters and11,
Sustainability 2019, mechanisms is illustrated in the following section.
x FOR PEER REVIEW 6 of 12

Figure 2. Pseudo code of the artificial bee colony (ABC) algorithm.

4. Numerical Experiment
The proper balance between diversification and intensification of the search process is critical to
find promising solutions
The test instances [22].designed
were In the ABC algorithm,
using this balance
the following mechanism.is accomplished by the algorithm’s
Both the non-hub nodes and
the hub nodes were randomly generated within a square range (e.g., 1000 × 1000 km ). Each nodeMore
parameter settings and control mechanism. The first parameter is the solution 2
number (SN). had
ainitial
certainsolutions
number indicate
of items to a be
more diversified
transported to asearch. The node.
destination second Theparameter is theamount
transportation abandonment
for each
criterion,
OD pair wasdenoted
randomlyas the limit. Awithin
generated higher abandonment
a range, criterion
such as [1,20] tons.means thattothe
Referring algorithm
reference [23],will
the
preserve
unit the current
transportation solution
cost was set by continuing
as 4.9 to digcurrency)
CNY (Chinese into its neighborhood
per ton-km. Previoussearch studies
area forsuggest
better
solutions.
that In thisvalue
a reasonable case, ofthe
thealgorithm
inter-hub performs as a discount
transportation more intensified search. Moreover,
factor is between 0.6 and 0.8, both the
and that
agreedy selection
reasonable valuemechanism in the neighborhood
of the non-hub-to-hub searchdiscount
transportation processfactor
and the probabilistic
is between 0.7 andselection of
0.9 [9,15].
solutions
In in the αonlooker
this research, bee phase
was initiated as 0.6, can β was initiated
and aggravate the intensified
as 0.8. The search. Theanalysis
sensitivity search for
performance
these two
using different
parameters combinations
is provided in theoffollowing
parameters and mechanisms
sections. is illustrated
First, we generated in instance
a test the following
with section.
20 nodes,
among which 4 hub nodes were predetermined. We labeled this instance as {1-4-20}, and we used this
4. Numerical
instance to find Experiment
the proper settings for the ABC algorithm and analyze the sensitivity of the proposed
model. After that,
The test instanceswe applied the proposed
were designed usingmodel and algorithm
the following to moreBoth
mechanism. complicated
the non-hubinstances
nodesso as
and
to
theexamine
hub nodes theirwere
effectiveness
randomly and efficiency.
generated The ABC
within algorithm
a square rangewas coded
(e.g., 1000 using
× 1000Javakmlanguage and
2). Each node

implemented on a personal computer with a 2.5-GHz processor. Each numerical

had a certain number of items to be transported to a destination node. The transportation amount for instance was executed
20 times
each OD repetitively, and thegenerated
pair was randomly averaged output
within awas usedsuch
range, to illustrate
as [1,20] the computational
tons. performance.
Referring to reference [23],
the unit transportation cost was set as 4.9 CNY (Chinese currency) per ton-km. Previous studies
4.1. Parameter Analysis
suggest that a reasonable value of the inter-hub transportation discount factor is between 0.6 and 0.8,
The acontrol
and that mechanism
reasonable value ofofthe
thenon-hub-to-hub
ABC algorithm is rather straightforward,
transportation as only
discount factor two parameters
is between 0.7 and
need to beIntuned
0.9 [9,15]. this research, 𝛼 was
(i.e., the size of initiated
bee colonyas 0.6, and𝛽the
(CS)and abandonment
was criterion
initiated as 0.8. (limit)). A
The sensitivity larger
analysis
setting
for these CS indicates
of two parameters a larger solutionin
is provided pool,
the which suggests
following a more
sections. diversified
First, searching
we generated paradigm,
a test instance
while
with 20 thenodes,
settingamong
of a higher
whichabandonment
4 hub nodes criterion limit implies aWe
were predetermined. more intensified
labeled searching
this instance scheme.
as {1-4-20},
and we used this instance to find the proper settings for the ABC algorithm and analyze the
sensitivity of the proposed model. After that, we applied the proposed model and algorithm to more
complicated instances so as to examine their effectiveness and efficiency. The ABC algorithm was
coded using Java language and implemented on a personal computer with a 2.5-GHz processor.
illustrate the computational performance.

4.1. Parameter Analysis

The control mechanism of the ABC algorithm is rather straightforward, as only two parameters
need to be 2019,
Sustainability tuned
11, (i.e.,
5197 the size of bee colony (𝐶𝑆) and the abandonment criterion (𝑙𝑖𝑚𝑖𝑡)). A larger 7 of 12
setting of 𝐶𝑆 indicates a larger solution pool, which suggests a more diversified searching
paradigm, while the setting of a higher abandonment criterion 𝑙𝑖𝑚𝑖𝑡 implies a more intensified
However, both
searching the setting
scheme. However, of aboth
largerthenumber
settingofofinitial solutions
a larger numberandofa higher abandonment
initial solutions and acriterion
higher
abandonment
would consumecriterion
much more would consume much
computational time. more computational
Therefore, we sought time. Therefore,
to balance we sought
the search to
capability
and the the
balance computational time.and
search capability Figures 3 and 4 illustrate
the computational time.the performance
Figures of the ABC
3 and 4 illustrate thealgorithm
performance for
of the ABC
solving the algorithm for solving
instance {1-4-20} with the instance
different {1-4-20}
settings with
of CS anddifferent
limit. Insettings of 𝐶𝑆 and
this research, 𝑙𝑖𝑚𝑖𝑡.
the size In this
of colony
research, the size of colony
and the abandonment and
criterion the set
were abandonment
as CS = 40 and criterion
limit = were
SN ∗setDim, 𝐶𝑆which
as in = 40 SN
andislimit = 𝑆𝑁 ∗
the number
𝐷𝑖𝑚,
of solutions
in which and𝑆𝑁Dim is the
is the dimensional
number lengthand
of solutions 𝐷𝑖𝑚 solution.
of each is the dimensional length of each solution.

390,000
390,000

385,000
385,000

380,000
380,000
cost
Transportation cost

375,000
375,000
Transportation

370,000
370,000 CS=20
CS=20
CS=40
CS=40
365,000
365,000 CS=80
CS=80
360,000
360,000

355,000
355,000

350,000
350,000
00 200
200 400
400 600
600 800
800 1000
1000 1200
1200
Iterations
Iterations

Figure
Figure
Figure 3.
3. 3. Algorithm
Algorithm
Algorithm performance
performance
performance with
with
with different
different
different settings
settings
settings of
of of 𝐶𝑆.
CS.𝐶𝑆.

390,000
390,000

385,000
385,000

380,000
380,000
cost
Transportation cost

375,000
375,000
Transportation

370,000
370,000

365,000
365,000

360,000
360,000

355,000
355,000

350,000
350,000
00 200
200 400
400 600
600 800
800 1000
1000 1200
1200
Iterations
Iterations
limit=0.5*SN*Dim
limit=0.5*SN*Dim limit=SN*Dim
limit=SN*Dim limit=2*SN*Dim
limit=2*SN*Dim

Figure
Figure 4.
4. 4.
Figure Algorithm
Algorithm performance
performance
Algorithm with
with
performance different
different
with settings
settings
different of
of of
settings 𝑙𝑖𝑚𝑖𝑡.
𝑙𝑖𝑚𝑖𝑡.
limit.

4.2. Sensitivity Analysis

4.2. Sensitivity Analysis
The determination of hubs is a critical factor affecting the entire air cargo transportation network.
The determination of hubs is a critical factor affecting the entire air cargo transportation
In the numerical
network. In the experiment, we assumedwe
numerical experiment, that each node
assumed had
that the node
each potential
hadtothe
be potential
a hub node.
to However,
be a hub
node. However, the upgrading from a non-hub node to a hub node is a strategic decision. Inonce
the upgrading from a non-hub node to a hub node is a strategic decision. In other words, othera
node is determined to be a hub node, its status will be maintained for a long period. Therefore, we
words, once a node is determined to be a hub node, its status will be maintained for a long period.
analyzed the
Therefore, we impact
analyzedof the
hubimpact
numbers on the
of hub transportation
numbers cost and distance.
on the transportation cost andWe also compared
distance. We also
the transportation cost and distance with two other scenarios (i.e., all direct transportation and all
compared the transportation cost and distance with two other scenarios (i.e., all direct transportation
transshipment transportation). As shown in Figures 5 and 6, the transportation cost and distance were
and all transshipment transportation). As shown in Figures 5 and 6, the transportation cost and
practically
distance werefixed practically
in the all-direct-transportation scenario, which is intuitively
fixed in the all-direct-transportation understandable
scenario, as all the
which is intuitively
transportation from the origin node to the destination node is transported directly without any stopover.
By comparison, the scenario of all transshipment transportation always possessed a larger transportation
distance than the scenario with all direct transportation. The transportation cost in the all transshipment
scenario was also higher than the all-direct-transportation scenario when the number of hubs was
small. However, with the increase of hub number, the transportation cost in the all-transshipment
scenario gradually decreased and finally became lower than the all-direct-transportation scenario even
though all transportation had to be transshipped.
always possessed a larger transportation distance than the scenario with all direct transportation.
The transportation cost in the all transshipment scenario was also higher than the
all-direct-transportation scenario when the number of hubs was small. However, with the increase
of hub number, the transportation cost in the all-transshipment scenario gradually decreased and
finally became
Sustainability lower
2019, 11, 5197 than the all-direct-transportation scenario even though all transportation8 had
of 12
to be transshipped.

700,000
700,000

650,000
650,000

600,000
600,000

cost
Transportation cost
550,000
550,000

Transportation
500,000
500,000

450,000
450,000

400,000
400,000

350,000
350,000

300,000
300,000
00 11 22 33 44 55 66 77 88 99
Number of
Number of hubs
hubs (p)
(p)
Optimal transportation
Optimal transportation All transshipment
All transshipment All direct
All direct transportation
transportation

Figure 5. The
The transportation
transportation cost
cost with
with different
different numbers of hubs.

15,000
15,000

14,000
14,000
13,000
13,000
distance
Transporation distance

12,000
12,000

11,000
11,000
Transporation

10,000
10,000
9000
9,000

8000
8,000
7000
7,000
6000
6,000
00 11 22 33 44 55 66 77 88 99
Number of
Number of hubs
hubs (p)
(p)
Optimal transportation
Optimal transportation All transshipment
All transshipment All direct
All direct transportation
transportation

Figure 6.
Figure 6. The
The transportation
transportation distance
distance with
with different
different numbers
numbers of
of hubs.

Comparatively
Comparativelyspeaking,
Comparatively speaking,Figure
speaking, Figure
Figure 5 reveals
55 revealsthatthat
reveals the optimal
that transportation
the optimal
the optimal solution
transportation
transportation had the had
solution
solution smallest
had the
the
transportation
smallest cost
transportationin comparison
cost in with the
comparison other
with two
the scenarios.
other two One thing
scenarios.
smallest transportation cost in comparison with the other two scenarios. One thing to notice is that to notice
One is
thing that
to the difference
notice is that
between
the the optimal
the difference
difference betweensolution
between the optimal
the and solution
optimal the all-transshipment
solution and the
and scenario became
the all-transshipment
all-transshipment scenario
scenariolessbecame
with the
became increasing
less
less with the
with the
number
increasing of hubs.
number Figure
of hubs.6 shows
Figure that
6 the
shows transportation
that the distance
transportation in the
distance
increasing number of hubs. Figure 6 shows that the transportation distance in the optimal solution optimal
in the solution
optimal slowly
solution
increased
slowly with thewith
slowly increased
increased increasing
with the numbernumber
the increasing
increasing of hubs,of
number and
of thenand
hubs,
hubs, reached
and a maximum
then reached
then reached distance,distance,
aa maximum
maximum after which
distance, it
after
after
began
which to
which it decrease.
it began
began to decrease.
to decrease.
Apart from the determination
Apart from the determination of of hubs,
hubs, thethe discount factor of
discount factor of inter-hub
inter-hub transportation
inter-hub transportation and
transportation and the
the
non-hub to hub transportation due to the economies of transportation
non-hub to hub transportation due to the economies of transportation scale also affect the route scale also affect the route
selection of
selection of the
of theair
the aircargo
air cargotransportation.
cargo transportation.As
transportation. Asshown
As shown
shown in in
Figure
in Figure7a,b,
Figure 7awe
7a and
and conducted
7b, we
7b, more numerical
we conducted
conducted more
more
experiments
numerical with different
experiments combinations
with different of α and β.
combinations Given of 𝛼
that
numerical experiments with different combinations of 𝛼 and 𝛽 . Given that the inter-hub the
and 𝛽
inter-hub
. Giventransportation
that the discount
inter-hub
factor was set asdiscount
transportation
transportation 0.6, the non-hub
discount factor was
factor to hub
was set as
set transportation
as 0.6, the
0.6, the non-hub discount
non-hub factor
to hub
to hub increased from
transportation
transportation 0.7 to 0.8
discount
discount and
factor
factor
from 0.8
increased to 0.9,
from the
0.7 total
to 0.8transportation
and from 0.8 cost
to increased
0.9, the total by 5.80% and
transportation 4.93%
increased from 0.7 to 0.8 and from 0.8 to 0.9, the total transportation cost increased by 5.80% and cost respectively,
increased by as shown
5.80% in
and
Figure 7a. When
4.93% respectively,
4.93% we
respectively, as fixed
as shownthe
shown in non-hub
in Figureto hub
Figure 7a. transportation
7a. When
When we we fixed discount
fixed the factor
the non-hub
non-hub to as 0.8
to hub and increased
hub transportation
transportation the
inter-hub transportation discount factor from 0.5 to 0.6 and from 0.6 to 0.7, the entire transportation
cost increased by 4.46% and 4.80% respectively, as shown in Figure 7b. In conclusion, the number
of hubs and their locations determine the structure of the air cargo transportation network, and the
determination of hubs and the transportation discount factor integrally affect the route selection for air
cargo transportation.
 discount factor as 0.8 and increased the inter-hub transportation discount factor from 0.5 to 0.6 and
from 0.6 to 0.7, the entire transportation cost increased by 4.46% and 4.80% respectively, as shown in
Figure 7b. In conclusion, the number of hubs and their locations determine the structure of the air
cargo transportation
Sustainability 2019, 11, 5197 network, and the determination of hubs and the transportation discount 9factor
of 12
integrally affect the route selection for air cargo transportation.

450,000 450,000

400,000 400,000

350,000 350,000

Transporation cost
300,000 300,000
Transportation cost

250,000 250,000

200,000 𝛼=0.6 200,000 𝛽=0.8

150,000 150,000

100,000 100,000

50,000 50,000

0 0
𝛽 = 0.7 𝛽 = 0.8 𝛽 = 0.9 𝛼 = 0.5 𝛼 = 0.6 𝛼 = 0.7

(a). The transportation cost change with different 𝛽 (b). The transportation cost change with different 𝛼

Figure
Figure 7.
7. The
The transportation
transportation cost
cost change
change with different settings of α𝛼and
with different andβ.𝛽.

4.3.
4.3.Algorithm
AlgorithmPerformance
Performance
In
Inorder
ordertotofurther
furtherexamine
examine and validate
and thethe
validate effectiveness
effectiveness andandefficiency of the
efficiency ofproposed
the proposedmodel and
model
algorithm, we generated more instances with different settings. As
and algorithm, we generated more instances with different settings. As shown in Table 1, 10shown in Table 1, 10 instances were
generated
instances withwerethe naming schema
generated with the as (instance
naming ID–hub
schema number–node number).number–node
as (instance ID–hub Also, for each number).
instance,
we compared the optimal solution with the scenarios of all direct transportation
Also, for each instance, we compared the optimal solution with the scenarios of all direct and all transshipment.
In Table 1, the column
transportation and alloftransshipment.
cost saving 1 was calculated
In Table 1, thebycolumn
comparisonof costof saving
the optimal
1 wastransportation
calculated by
cost and the all-direct-transportation cost, while the column of cost
comparison of the optimal transportation cost and the all-direct-transportation cost, while saving 2 was calculated by the
the
comparison of the optimal transportation cost with the all-transshipment-transportation
column of cost saving 2 was calculated by the comparison of the optimal transportation cost with cost. Fromthe
Table 1, for all instances, the optimalcost.
all-transshipment-transportation solution
FromcanTable
provide 1, average transportation
for all instances, the cost savings
optimal of 16.42%
solution can
and 13.89% in contrast to the all-direct-transportation scenario
provide average transportation cost savings of 16.42% and 13.89% in contrast to the and the all-transshipment scenario
respectively. For the instances
all-direct-transportation withand
scenario same thenode number, the increase
all-transshipment scenario of hub number in
respectively. Fora certain range
the instances
could
with reduce
same node the optimal
number, transportation
the increase costof hubandnumber
the all-transshipment-transportation
in a certain range could reduce cost.the
Moreover,
optimal
ittransportation
is a popular trend to employ transshipment transportation along with the increase
cost and the all-transshipment-transportation cost. Moreover, it is a popular trend to of the network
complexity and the number
employ transshipment of hubs. Fromalong
transportation instance
with 4 tothe
instance 10, the
increase all-transshipment-transportation
of the network complexity and the
cost became lower than that of the all-direct-transportation mode,
number of hubs. From instance 4 to instance 10, the all-transshipment-transportation and this differentiation became even
cost became
larger. Regarding the transportation distance, as shown in Table 1, the optimal
lower than that of the all-direct-transportation mode, and this differentiation became even larger. solution generated 8.33%
more distance
Regarding theon average in contrast
transportation to the
distance, as all-direct-transportation
shown in Table 1, the optimal mode. solution
However,generated
in comparison
8.33%
with the all-transshipment mode, the optimal solution could actually reduce
more distance on average in contrast to the all-direct-transportation mode. However, in comparison the transportation distance
by 21.65%
with the on average. In addition,
all-transshipment mode,increasing
the optimalnetwork
solutioncomplexity and hub numbers
could actually reduce the could reduce the
transportation
extra transportation
distance by 21.65% on distance
average. dueIntoaddition,
the transshipment
increasingarrangements.
network complexity and hub numbers could
reduce the extra transportation distance due to the transshipment arrangements.
Sustainability 2019, 11, 5197 10 of 12

Table 1. The comparison of transportation cost and distance with different instances. The Instance naming format is (instance ID–hub number–node number).

Transportation Cost (CNY) Transportation Distance (km)

Distance Distance
Instance Scenario A Scenario B Scenario C Cost Saving 1 Cost Saving 2 Scenario A Scenario B Scenario C
Increase Decrease
(1–2–20) 438,713.91 390,969.50 583,109.91 10.88% 32.95% 8052.38 8550.10 13,462.65 6.18% 36.49%
(2–4–20) 438,713.91 365,149.01 483,464.52 16.77% 24.47% 8052.38 8427.06 11,934.45 4.65% 29.39%
(3–4–30) 885,976.78 751,349.91 900,557.43 15.20% 16.57% 13,990.30 14,966.31 20,446.80 6.98% 26.80%
(4–5–30) 885,976.78 730,280.24 833,628.52 17.57% 12.40% 13,990.30 15,076.91 19,424.46 7.77% 22.38%
(5–5–40) 967,544.72 808,134.01 927,595.89 16.48% 12.88% 17,942.00 19,564.89 24,579.63 9.05% 20.40%
(6–6–40) 967,544.72 799,572.63 897,651.72 17.36% 10.93% 17,942.00 19,683.74 24,457.99 9.71% 19.52%
(7–6–60) 1,442,330.17 1,217,137.02 1,341,243.04 15.61% 9.25% 29,271.50 32,091.78 39,037.06 9.63% 17.79%
(8–8–60) 1,442,330.17 1,183,377.64 1,276,808.14 17.95% 7.32% 29,271.50 32,135.35 37,950.30 9.78% 15.32%
(9–8–80) 1,604,880.79 1,329,213.29 1,425,454.36 17.18% 6.75% 34,978.61 38,268.57 45,217.90 9.41% 15.37%
(10–10–80) 1,604,880.79 1,296,770.24 1,371,033.86 19.20% 5.42% 34,978.61 38,518.37 44,278.34 10.12% 13.01%
Average 16.42% 13.89% 8.33% 21.65%
Sustainability 2019, 11, 5197 11 of 12

5. Conclusions
Air cargo transportation has been taking an increasingly important role in transportation industry,
which requires different network design and route selection in contrast to air passenger transportation.
In this research, we proposed an integrated model simultaneously considering air cargo transportation
network design and flight route selection. The ABC algorithm was adapted to solve the proposed
model with a succinct solution representation scheme and proper combinations of parameters and
control mechanisms. The numerical experiment suggests that the determination of hubs and the
selection of transshipment mode is core to air cargo transportation management. This research can
provide managerial insights for the development of air cargo transportation.
There are some limitations to this research. One is the lack of consideration regarding the extra
time and cost required due to consolidation and transshipment operations, which may weaken the
solution effect. Another limitation is that the parameter-value settings are drawn from previous
literature; it would be more convincing if practical operational data could be collected and analyzed.
Therefore, our future research direction is to investigate more practical data and conduct some real
case studies applying the proposed model.

Author Contributions: Conceptualization, C.K.M.L.; methodology, K.K.H.N. and S.Z.; software, K.K.H.N. and
S.Z.; validation, K.K.H.N. and S.Z.; writing—original draft preparation, K.K.H.N. and S.Z.; writing—review and
editing, C.K.M.L.
Funding: This work was funded in part by the National Natural Science Foundation of China under grant number
71902171, in part by the Humanities and Social Science Foundation of Ministry of China under grant number
19YJC630216, in part by the Natural Science Foundation of Zhejiang Province, China under the grant number
LY19G010003, and supported by a grant from a collaborative research project of the Department of Industrial and
Systems Engineering, the Hong Kong Polytechnic University (Project No. K-ZM2K).
Conflicts of Interest: The authors declare no conflicts of interest.

References
1. Merkert, R.; Van de Voorde, E.; de Wit, J. Making or breaking-Key success factors in the air cargo market.
J. Air Transp. Manag. 2017, 61, 1–5. [CrossRef]
2. IATA. Annual Review. Available online: https://www.iata.org/publications/Pages/annual-review.aspx,
(accessed on 10 September 2019).
3. Suwanwong, T.; Sopadang, A.; Hanaoka, S.; Rodbundith, T. Evaluation of air cargo connectivity and policy
in Thailand. Transp. Policy 2018, 72, 24–33. [CrossRef]
4. Hong, S.J.; Randall, W.; Han, K.; Malhan, A.S. Estimation viability of dedicated freighter aircraft of
combination carriers: A data envelopment and principal component analysis. Int. J. Prod. Econ. 2018, 202,
12–20. [CrossRef]
5. Lange, A. Does cargo matter? The impact of air cargo operations on departure on-time performance for
combination carriers. Transp. Res. Part A Policy Pract. 2019, 119, 214–223. [CrossRef]
6. Kim, H.K.; Lee, C.W. Development of a Cost Forecasting Model for Air Cargo Service Delay Due to Low
Visibility. Sustainability 2019, 11, 4390. [CrossRef]
7. Kasilingam, R.G. Air cargo revenue management: Characteristics and complexities. Eur. J. Oper. Res. 1997,
96, 36–44. [CrossRef]
8. Yan, S.; Chen, S.C.; Chen, C.H. Air cargo fleet routing and timetable setting with multiple on-time demands.
Transp. Res. Part E Logist. Transp. Rev. 2006, 42, 409–430. [CrossRef]
9. Aykin, T. Networking Policies for Hub-and-Spoke Systems with Application to the Air Transportation
System. Transp. Sci. 1995, 29, 201–221. [CrossRef]
10. Yang, T.H. Stochastic air freight hub location and flight routes planning. Appl. Math. Model. 2009, 33,
4424–4430. [CrossRef]
11. Lin, C.C.; Lin, Y.J.; Lin, D.Y. The economic effects of center-to-center directs on hub-and spoke networks for
air express common carriers. J. Air Transp. Manag. 2003, 9, 255–265. [CrossRef]
12. Lin, C.C.; Chen, S.H. An integral constrained generalized hub-and-spoke network design problem. Transp.
Res. Part E Logist. Transp. Rev. 2008, 44, 986–1003. [CrossRef]
Sustainability 2019, 11, 5197 12 of 12

13. Tan, P.Z.; Kara, B.Y. A hub covering model for cargo delivery systems. Networks 2005, 49, 28–39. [CrossRef]
14. Oktal, H.; Ozger, A. Hub location in air cargo transportation: A case study. J. Air Transp. Manag. 2013, 27,
1–4. [CrossRef]
15. O’Kelly, M.E.; Bryan, D.L.; Skorinkapov, D.; Skorinkapov, J. Hub network design with single and multiple
allocation: A computational study. Locat. Sci. 1996, 4, 125–138. [CrossRef]
16. O’Kelly, M.E.; Bryan, D. Hub location with flow economies of scale. Transp. Res. Part B Methodol. 1998, 32,
605–616. [CrossRef]
17. Bryan, D.L.; O’Kelly, M.E. Hub and spoke networks in air transportation: An analytical review. J. Reg. Sci.
1999, 39, 275–295. [CrossRef]
18. Kim, H. p-Hub protection models for survivable hub network design. J. Geogr. Syst. 2012, 14, 437–461.
[CrossRef]
19. Ohashi, H.; Kim, T.; Oum, T.H.; Yu, C. Choice of Air Cargo Transshipment Airport: An Application to Air
Cargo Traffic to/from Northeast Asia. J. Air Transp. Manag. 2005, 11, 149–159. [CrossRef]
20. Kim, J.Y.; Park, Y. Connectivity analysis of transshipments at a cargo hub airport. J. Air Transp. Manag. 2012,
18, 12–15. [CrossRef]
21. Karaboga, D.; Basturk, B. A powerful and efficient algorithm for numerical function optimization: artificial
bee colony (ABC) algorithm. J. Glob. Optim. 2007, 39, 459–471. [CrossRef]
22. Zhang, S.Z.; Lee, C.K.M.; Yu, K.M.; Lau, H.C.W. Design and development of a unified framework towards
swarm intelligence. Artif. Intell. Rev. 2016, 47, 253–277. [CrossRef]
23. Lin, C.C.; Lin, J.Y.; Chen, Y.C. The capacitated p-hub median problem with integral constraints: An application
to a Chinese air cargo network. Appl. Math. Model. 2012, 36, 2777–2787. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
 © 2019. This work is licensed under
http://creativecommons.org/licenses/by/3.0/ (the “License”). Notwithstanding
the ProQuest Terms and Conditions, you may use this content in accordance
with the terms of the License.