The signing of the Regional Comprehensive Economic Partnership agreement brings new opportunities for the development of international air transportation. Faced with fierce competition, it is worth studying how hub airports should enhance competitiveness, and how low-cost carriers and full-service carriers should optimize the RCEP international airline network layout for better development. Aiming at providing suggestions for the development of hub airports, low-cost and full-service carriers in the RCEP international airline network, this paper identifies the hub airports, analyzes the layout of the RCEP international airline network, and the multi-layered characteristics based on an improved contribution matrix using data from 2010 to 2019 collected from the Official Airline Guide (OAG). This method comprehensively considers attributes of hub airports and the multi-layered characteristics of the airports and routes. The layout analysis indicates that the RCEP international transportation market presents a more open environment for competition and cooperation where base carriers are often the biggest supporters of hub construction. The multi-layered characteristics analysis reveals that low-cost carriers contribute more towards opening up new RCEP routes than full-service carriers. It is advised that carriers newly entering the RCEP international aviation transportation market and low-cost carriers dedicate to establishing new routes around their hub airports to monopolize this market and enhance their market share, whilst full-service carriers consolidate existing routes and increase route density to achieve economic benefits.
Air transportation has recently become an important way of carrying passengers and cargo within regional economies due to its distinctive benefits of high levels of internationalization and excellent long-distance accessibility[
The signing of the Regional Comprehensive Economic Partnership (RCEP) agreement by ten ASEAN countries in 2020 has created new opportunities for the growth of carriers and international air transportation in these countries. On the one hand, the competitiveness of the hub airports and the overall layout of the RCEP international airline network need to be explored. On the other hand, given the fact that the market competition over carriers tends to be more intense with the expansion of the RCEP international airline network and the increasing number of carriers, it is worth studying how low-cost and full-service carriers should optimize the international RCEP airline network layout to improve competitiveness. Considering the traffic interaction between airports, the balance between safety and efficiency, and the coordinated interactive behavior of passengers and flights under big data[
With the purpose of providing suggestions for the development of hub airports, low-cost and full-service carriers in the RCEP international airline network, this paper evaluates the importance of airports, identifies the hub airports, analyzes the overall layout of the RCEP international airline network, and comparatively analyzes carriers' network layouts and multi-layered characteristics based on an improved contribution matrix using data from 2010 to 2019 collected from the Official Airline Guide (OAG). Evaluating the importance of airports and identifying the regional hubs of the RCEP international airline network is helpful for evaluating the competitiveness of hub airports and analyzing the network layout in the RCEP region. The comparative analysis of network layouts and multi-layered characteristics can provide references for low-cost carriers and full-service carriers to optimize their RCEP international airline network layouts and improve their competitiveness.
The main contributions of this paper are as follows:
- This paper proposes a method for evaluating airport nodes' importance and identifying hub airports in airline networks based on an improved contribution matrix, which comprehensively considers the weight of the network, the contribution of neighboring airports, and the multi-layered characteristics of airports compared to previous studies. The nodes are classified into four categories and the Susceptible-Infectious-Recovered (SIR) model is employed to assess the effectiveness of the proposed method and two other methods.
- The competitiveness of hub airports and the overall layout of the RCEP international airline network are explored from 2010 to 2019. The layout of carriers' international networks in the RCEP region and the multi-layered characteristics are further comparatively analyzed based on the multi-layered metrics.
- Based on the analysis of the overall and carriers' RCEP international airline network layout, and the multi-layered characteristics of airports and routes, suggestions for the development and optimization of hub airports, full-service carriers, and low-cost carriers in the RCEP international airline network are further provided.
The remainder of the paper is organized as follows. Section "Related works" reviews the most relevant paper about the two main methods of accessing the importance of airport nodes. Section "Methods and model" establishes a multi-layer regional airline network of different carriers based on the theory of multi-layer complex network, proposes an influential node identification method based on the improved contribution matrix, introducing the SIR model for validating method effectiveness, and multi-layer metrics used for analyzing the multi-layered characteristics of airports and routes. Section "Example analysis" accesses the importance of the airport nodes, classifies the airports into four categories, and verifies the rationality of the proposed methodology by SIR model. In the section "Layout analysis of the RCEP international airline network", we identify the hubs, classify the airport nodes, analyze the layout of the RCEP international airline networks from overall and typical carriers' perspectives, and explore the multi-layered characteristics of airports and routes. Section "Suggestions for airports and carriers" further provides suggestions for the development of the RCEP international airline network from the perspectives of airports and carriers. Finally, the results are concluded in the section "Conclusion".
Regional airline network hubs usually refer to important airports in a geographical area that connect different aviation markets, have convenient transit, and are densely populated with flights. Airline network layout refers to the coverage of routes, frequency, connectivity, and the selection of hub airports of carriers.
To analyze the airline network layout, previous studies have assessed the importance of airport nodes and classified airports. The methods of hub airport importance assessment and identification of previous studies are mainly classified into two categories: first, the classification of airport nodes using descriptive statistics and clustering models. Clustering models in airline network research groups or clusters nodes into subsets with similar characteristics or functions, which can help reveal the internal structure of the airline network, the patterns of interactions between nodes, and potential functional modules. Fuellhart et al. classified 131 airports in 53 global multiple airport regions and analyzed the role of these airports using hierarchical cluster analysis, utilizing supply-side variables such as aspects of competition, routes, and aircraft capacities[
Since the regional airline network is a typical complex network, another useful way of identifying aviation hubs is by evaluating the influence and importance of airport nodes in airline networks using complex network theory. Zhou et al. proposed a method for determining key nodes in complex networks using importance evaluation matrices by defining node efficiency and node importance evaluation matrices[
Table 1 summarizes the literature in terms of network type, network weight, and whether these papers considered the multi-layered characteristics of the network and the impact of neighboring nodes. Most studies focused on regional networks such as a certain country or continent, whilst the research on carrier networks nearly lacked investigation. In addition, most studies concentrated on the unweighted airline network structure, whereas the actual network is weighted. The weight is usually the number of flights and it impacts the accurate identification and classification of hub airports. Furthermore, few studies have considered the multi-layered characteristics of networks. Constructing a multi-layer airline network based on carriers can reveal the deeper internal relationships among them because a regional airline network consists of networks used by different carriers, which can provide carriers with more references when selecting airports, designing routes, and optimizing airline network layout[
Table 1 Review of related works.
Ref Network type Network weight Multi-layered characteristics Impact of neighboring nodes Regional network Carrier network Weighted Unweighted Fuellhart et al. (2018) ✓ ✓ Wang et al. (2019) ✓ ✓ Rotondo (2019) ✓ ✓ Gao (2020) ✓ ✓ Kim et al. (2021) ✓ ✓ Chen et al. (2022) ✓ ✓ Zhou et al. (2012) ✓ ✓ ✓ Verma et al. (2014) ✓ ✓ ✓ Du et al. (2016) ✓ ✓ ✓ Tu et al. (2017) ✓ ✓ ✓ Wong et al. (2019) ✓ ✓ ✓ ✓ Xu et al. (2021) ✓ ✓ Du et al. (2023) ✓ ✓ Yang et al. (2024) ✓ ✓ ✓ This paper ✓ ✓ ✓ ✓ ✓
This paper constructs weighted airline networks, identifies the hub airports, and analyzes the overall and the carriers' RCEP international airline networks based on an improved contribution matrix, which considers the contribution of neighboring airports and the multi-layered characteristics of the airports and routes. At the same time, it analyzes the multi-layered characteristics of RCEP international airline networks to provide suggestions for the development of hub airports, full-service carriers, and low-cost carriers.
This section describes the models and metrics utilized for the analysis in this paper. Initially, the foundation of hub identification and airline network layout analysis—the construction of a multi-layer regional airline network is presented. Subsequently, the foundation of the influential node identification method is expounded, including the basic topological metrics and contribution matrix. Following that, the Susceptible-Infectious-Recovered (SIR) model is introduced, which is employed to verify the rationality of the proposed method. Lastly, the multi-layer metrics employed for the analysis of the multi-layered characteristics of airports and routes are displayed. On this basis, the overall layout of the RCEP international airline network is analyzed, and the comparative analysis of carriers' RCEP international airline network layout and its multi-layered characteristics is conducted to provide insights for the development of airports, full-service carriers, and low-cost carriers in the RCEP region. The mind map of this paper is shown in Fig. 1.
Graph: Figure 1 Mind map of this paper.
The construction of the multi-layer regional airline networks is the basis of the proposal of an influential airport identification method and the layout analysis of the RCEP international airline network. This paper mainly considers the direction and frequency of the routes and constructs a directed weighted regional airline network.
We first establish the single-layer directed weighted airline network of each carrier in the region, which uses the airport as the network node, the route connection as the directed edge, and the number of flights of the route during the observation period as the edge weight. Airline network is expressed as
We then establish the regional multi-layer airline network based on the single-layer airline network of the carriers according to the multi-layer complex network theory[
The mapping approach enables the aggregation of single-layer airline networks of carriers in a regional airline network, thus simplifying the representation and analysis of multi-layer regional airline networks[
To identify the hub airports of a regional airline network, we comprehensively discuss the basic topological metrics of airport nodes and propose an influential node identification method based on the improved contribution matrix, which considers the contribution of neighboring airports and the multi-layered characteristics of the airports. The Susceptible-Infectious-Recovered (SIR) model is introduced to verify the rationality of the proposed method.
There are several basic topological metrics of nodes, which are used to explore the influence and importance of airport nodes in airline networks. The influence of airport nodes can be practically evaluated from both local and global perspectives. The local influence of the airport is mostly reflected in its transportation scale, which includes the number of airport routes and flights and can be represented by the indexes of node degree
Node degree
Graph
Node strength
2
Graph
Betweenness centrality
3
Graph
where
Previous studies have found that the contribution matrix can be applied to identify key airport nodes in undirected airline networks by integrating airport location, route flow and other factors[
4
Graph
where
However, the traditional contribution matrix only considers the node degree from a local perspective, causing a biased assessment of airport nodes' influence. To make the evaluation more comprehensive, we improve the traditional contribution matrix by combining the metrics of node strength and betweenness centrality to weigh the important contribution proportion of airport nodes, which can maintain the directional characteristics of routes. The improved contribution matrix of airport nodes in the airline network is therefore expressed as:
5
Graph
where
Airport importance
6
Graph
where the contribution from the airport itself is defined as
Airport importance
The Susceptible-Infectious-Recovered (SIR) model has been extensively employed to assess the effectiveness of the node importance evaluation method and various centrality measures. The model classifies nodes into three categories: susceptible nodes, infected nodes, and immune nodes. At moment t, the number of susceptible nodes is denoted as
Graph
The influence of the initial set of infected nodes is measured by the number of the immune nodes
Since the regional multi-layer airline network consists of routes used by different carriers, it is necessary to consider multi-layer metrics to analyze the attributes of airport nodes. This paper uses multi-layer complex network metrics, including node activity, multiplex participation coefficient, and route uniqueness to further explore the influence of the multi-layered characteristics of airports and routes and explore the relationship between carriers and airports.
In a multi-layer network, node
7
Graph
where
A high node degree of an airport in the single-layer airline network of a carrier suggests that the airport is vital for this carrier. However, it is worth noting that the importance of an airport can vary between different carriers. Multiplex participation coefficient
8
Graph
where
The airline networks of different carriers may contain the same airport nodes, but the route layouts may be quite different due to the influence of the business model, transportation market, or air traffic rights policy. Route uniqueness indicates the proportion of routes in each carrier's single-layer airline network that do not appear in other layers of the regional multi-layer airline network, reflecting the difference and uniqueness of each carrier's route layout in the regional multi-layer airline network:
9
Graph
where
This section mainly assesses the node importance and classifies the nodes of an example network based on the proposed node importance assessment method, and uses the SIR model to compare the method proposed in this paper with the node importance assessment method in references 14 and 17 to validate the reasonableness of the proposed method, which lays the foundation for identifying the hubs of the RCEP international airline network and analyzing the network layout in the following section.
Routes and flights usually exist in both directions between two airports. We can simplify the calculation by taking a randomly generated weighted topological network as an example of airport node evaluation and classification while ignoring the difference of round-trip flights and directly providing the edge weights, as shown in Fig. 2.
Graph: Figure 2 A randomly generated weighted topological network.
The airport node evaluation method of the airline network based on the improved contribution matrix is used to calculate the importance of each node. The results and rankings are listed in Table 2.
Table 2 Airport importance of the weighted topological network.
No. Node Proportion of 1 v11 1.000 0.129 1.129 88.6 2 v14 0.183 0.812 0.994 18.4 3 v7 0.099 0.746 0.845 11.8 4 v10 0.000 0.772 0.772 0.0 5 v13 0.023 0.748 0.771 2.9 6 v12 0.000 0.746 0.746 0.0 7 v4 0.198 0.133 0.331 59.8 8 v5 0.104 0.129 0.233 44.7 9 v6 0.047 0.089 0.136 34.6 10 v3 0.027 0.100 0.128 21.4 11 v9 0.107 0.005 0.112 95.3 12 v16 0.052 0.052 0.104 49.9 13 v15 0.000 0.102 0.102 0.0 14 v8 0.000 0.090 0.090 0.0 15 v1 0.018 0.061 0.079 22.9 16 v2 0.049 0.013 0.062 78.9
Based on the difference between the importance of the airport and the proportion of its contribution, the airport nodes can be categorized into four categories according to the definition of airport categories in the hub-and-spoke network:
This section first calculates the node importance of the example network using the proposed method, the method in references 14 and 17 and analyzes the difference in the ranking results. Furthermore, we compare and analyze the reasonableness of the proposed method using the SIR model.
The rankings of airport importance differ slightly because each method focuses on different aspects when assessing node importance. The airport importance calculation results and rankings of different methods are listed in Table 3. WD, WI and WP are the methods in this paper, reference 14 and 17 respectively. For example, v2 and v3 are similar in edge weight and location. The other two methods emphasize the difference in edge weights, which correspond to the flight frequency, leading to that v2 is more significant than v3. However, the proposed method suggests that carriers will consider the hub role of airport location the priority when choosing a new airport, rather than the flight frequency. When there is a small difference in edge weights between airports, v3 adjacent to local-hub airport v4 is more advantageous than v2 adjacent to feeder-line airport v1 for the carriers. Besides, v16 ranks high in the other two methods but low in this paper. Although v16 is identified as a local-hub airport, its importance is relatively low in the airline network since it is not in the center of the airline network and neither its airport scale nor distribution function has a substantial impact.
Table 3 Node importance ranking results for different methods.
No. WD in this paper WI in reference 14 WP in reference 17 Airport node Airport node Airport node 1 v11 1.129 v11 0.412 v11 0.622 2 v14 0.994 v14 0.398 v14 0.512 3 v7 0.845 v16 0.378 v16 0.498 4 v10 0.772 v7 0.376 v7 0.445 5 v13 0.771 v12 0.366 v4 0.425 6 v12 0.746 v15 0.356 v12 0.402 7 v4 0.331 v4 0.355 v9 0.395 8 v5 0.233 v13 0.345 v15 0.394 9 v6 0.136 v2 0.342 v10 0.377 10 v3 0.128 v3 0.338 v13 0.374 11 v9 0.112 v10 0.322 v6 0.364 12 v16 0.104 v6 0.315 v5 0.355 13 v15 0.102 v5 0.295 v2 0.298 14 v8 0.090 v1 0.266 v1 0.256 15 v1 0.079 v9 0.242 v8 0.197 16 v2 0.062 v8 0.195 v3 0.128
The nodes in the top 3, 4, 5, and 6 of the importance ranking results of each method in Table 3 are used as the initial infected node set, respectively. The SIR model is run 100 times and the number of iterations is taken as 30, so as to compare the node influence of different initial infected node sets and the effectiveness of different methods. The faster the number of immune nodes increases, the higher the method's effectiveness. The results are shown in Fig. 3, with the horizontal axis denoting the number of iterations and the vertical axis denoting the node influence.
Graph: Figure 3 Node influence of different methods, different initial set of infected nodes.
When the initial set of infected nodes contains 3 nodes, the node influence of WI rises faster than WD and WP when the number of iterations is 1. However, the influence of the set of nodes identified by WD is higher than that of the WP and WI methods overall. The same conclusion can be drawn when there are 4, 5, 6 nodes in the initial set of infected nodes. Therefore the effectiveness of the node influence assessment method proposed in this paper is comparatively verified. In Fig. 3a, WD needs 8 iterations to immunize all nodes, while in Fig. 3d, it can immunize all nodes after 6 iterations. This is attributed to the fact that the initial set of infected nodes in Fig. 3d contains 6 nodes, which can infect and immunize other nodes in the network more quickly.
This section identifies the hubs and classifies the airport nodes of the RCEP international airline network from 2010–2019 based on the proposed methodology. On this basis, the overall layout of the RCEP international airline network during the years 2010–2019 is analyzed and the comparative analysis of carriers' RCEP international airline network layout and its multi-layered characteristics is conducted, which provides insights for the development of airports, full-service carriers, and low-cost carriers in the RCEP region in the following section.
The data analyzed in this paper is obtained from the Official Airline Guide database (https://oag.cn/). We first collect the data on international flights in the RCEP region during 2010–2019 and construct the RCEP multi-layer international airline network. We then apply the methods to identify the regional aviation hubs of the RCEP and the multi-layer attributes of airport nodes. We make the following assumptions to better organize the analysis:
- The regional routes from China to Hong Kong and Macao are considered international routes, but international routes for Taiwan are not considered.
- The origin-to-destination and destination-to-origin directions are considered two routes.
- Routes with stops are broken up into multiple routes comprising origin-to-stop and stop-to-destination legs since they involve one or more stop airports in addition to the arrival and departure airport.
The top ten most important airports in the RCEP international airline network in 2019 are listed in Table 4. Airports including SIN, ICN, HKG, KUL, BKK, and PVG are found to be the core-hub airports of the RCEP international airline network since they all have the characteristics of high airport importance (
Table 4 Importance of the top ten airports in the RCEP international airline network in 2019.
No. Airport node Proportion of 1 SIN (Singapore Changi International Airport) 1.000 1.019 2.019 49.5 2 ICN (Incheon International Airport) 0.551 1.316 1.867 29.5 3 HKG (Hong Kong International Airport) 0.436 1.364 1.800 24.2 4 KUL (Kuala Lumpur International Airport) 0.312 1.481 1.793 17.4 5 BKK (Bangkok International Airport) 0.350 1.435 1.785 19.6 6 PVG (Shanghai Pudong International Airport) 0.115 1.577 1.692 6.8 7 KIX (Japan Kansai International Airport) 0.033 1.595 1.629 2.05 8 CAN (Guangzhou Baiyun International Airport) 0.013 1.612 1.626 0.82 9 PEK (Beijing Capital International Airport) 0.008 1.615 1.623 0.47 10 SYD (Sydney International Airport) 0.031 1.591 1.621 1.89
As can be seen in Fig. 4, the airports of the RCEP international airline network are divided into four categories according to the difference between the importance of the nodes and the proportion of the contribution. It is found that with the growth of the total number of airport nodes in the past 10 years, the number of core-hub airports has remained stable at six, whereas the numbers of key-spoke airports and local-hub airports have increased significantly and that of feeder-line airports has decreased slowly. It can be inferred that the core-hub airports of the RCEP international airline network are very stable, and the regional airline network is constantly expanding with the number of key-spoke airports around the core-hub airport increasing, which reflects the remarkable nature of the hub-and-spoke structure characteristics of the network.
Graph: Figure 4 Classification of airport nodes in the RCEP international airline network from 2010 to 2019.
As indicated in Table 5, we further research the airport development and layout evolution of the RCEP international airline network during 2010–2019. Table 5 has been sorted in descending order of node importance and lists the node contribution of itself. Although constituting a small proportion of the contribution, airports including BKK, KUL and PVG have been the top ten most important airports in the RCEP international airline network since 2010. The proportions of contribution made by airports BKK, KUL and PVG have shown an overall increasing trend, indicating that their own airport scale and hub role are constantly increasing, and their influence on the overall network is remarkable. In contrast, the proportions of the contribution made by airports PEK and SYD have fluctuated slightly over the past 10 years. These airports mostly rely on the influence of neighboring ones to maintain their importance, indicating that airports PEK and SYD have always been key-spoke airports of the RCEP international airline network. Although airports SIN, ICN and HKG were not among the ten most important airports in 2010 and 2013, they still rank high (
Table 5 Airport importance ranking top 25 in the RCEP international airline network from 2010 to 2019.
No. 2010 2013 2016 2019 Airport node Proportion of Airport node Proportion of Airport node Proportion of Airport node Proportion of 1 BKK 4.72 KUL 8.42 SIN 37.72 SIN 49.52 2 KUL 7.70 PVG 5.02 ICN 29.68 ICN 29.50 3 PVG 4.22 BKK 3.32 HKG 28.87 HKG 24.23 4 NRT 0.80 PEK 0.60 KUL 12.28 KUL 17.41 5 SYD 1.29 MNL 0.25 PVG 10.59 BKK 19.63 6 MNL 0.49 SYD 0.76 BKK 11.17 PVG 6.81 7 CAN 0.71 CAN 0.37 PEK 1.50 KIX 2.05 8 PEK 0.94 NRT 0.41 KIX 2.17 CAN 0.82 9 SGN 0.38 CGK 0.60 SGN 0.48 PEK 0.47 10 CGK 0.39 SGN 0.21 DPS 0.91 SYD 1.89 11 KIX 0.85 KIX 0.45 MNL 0.20 MNL 0.11 12 HKT 0.04 HKG 24.16 HAN 0.06 DPS 0.91 13 DPS 0.12 DPS 0.30 NRT 0.57 HAN 0.19 14 AKL 0.10 HAN 0.07 CGK 0.06 SGN 0.34 15 HAN 0.17 ICN 23.01 SYD 0.62 REP 1.88 16 SIN 28.83 MEL 0.26 CTU 0.06 CKG 0.09 17 MEL 0.29 HKT 0.02 REP 0.17 NRT 0.70 18 HKG 23.26 AKL 0.05 KMG 0.76 XIY 0.04 19 KMG 0.10 CTU 0.00 AKL 0.47 CTU 0.20 20 PNH 0.02 KMG 0.26 XIY 0.03 AKL 1.07 21 XMN 0.01 PNH 0.01 CSX 0.03 DAD 0.09 22 BNE 0.26 SIN 37.17 CKG 0.02 HKT 0.38 23 ICN 25.83 XMN 0.01 CNX 0.03 PNH 0.76 24 CRK 0.00 BNE 0.12 HKT 0.15 KMG 0.42 25 BKI 0.02 RGN 0.01 HND 0.11 XMN 0.02
The evolution of the core-hub airports also reflects the evolution of the RCEP international airline network layout. In 2010, the airline network mostly focused on the core-hub airports SIN, HKG and ICN with significant hub-and-spoke characteristics in the local area, and point-to-point structure characteristics in the overall area. But with the rapid development of the core-hub and key-spoke airports, the influence on the other airports has continuously increases. The global influence of the core-hub airports tended to be stable after 2016, which indicates that the RCEP international airline network is a stable multi-distribution and non-strict hub-and-spoke airline network.
In the past decade, the RCEP international airline network size of various airlines has been continuously expanding, and the number of routes has significantly increased. The number of carriers with more than 100 routes has increased from 5 in 2010 to 24 in 2019. The top 20 carriers in terms of airline network size each year are shown in Table 6. Among the carriers with stopover routes in 2010, with the increase in the number of routes, the proportion of major carriers with stopover routes generally decreased significantly from 2016 to 2019, such as China Southern Airlines (CZ) and China Eastern Airlines (MU). Only a few carriers have seen an increase in the proportion of stopover routes, such as Spring Airlines (9C), which accounted for 18.6% of stopover routes in 2019, significantly higher than traditional full-service carriers. In 2010, carriers without non-stop routes added direct routes in the following ten years, such as Korean Airlines (KE) and Asian Airlines (OZ). Based on this, four typical carriers are selected to analyze their layouts in the RCEP international airline network: MU, the full-service carrier with stopover routes, OZ, the full-service carrier without stopover routes, 9C, low-cost carrier with stopover routes, and AK (Air Asia), low-cost carrier without stopover routes.
Table 6 Size of RCEP international airline networks of different carriers from 2010 to 2019 (top 20).
Rank 2010 2013 2016 2019 Carrier Routes number Stopover routes proportion (%) Carrier Routes number Stopover routes proportion (%) Carrier Routes number Stopover routes proportion (%) Carrier Routes number Stopover routes proportion (%) 1 CZ 160 17.50 MU 221 20.81 MU 326 23.31 MU 375 13.87 2 MU 154 16.88 KE 160 0.00 CZ 254 15.75 CA 310 9.03 3 KE 146 0.00 CZ 141 18.44 KE 148 0.00 CZ 294 6.12 4 OZ 124 0.00 OZ 138 0.00 OZ 144 0.00 SQ 229 7.86 5 CA 112 30.36 CA 103 24.27 CA 142 14.08 NH 212 3.77 6 AK 92 0.00 AK 100 0.00 AK 128 0.00 JL 206 5.83 7 MH 91 18.68 VN 96 12.50 VN 122 4.92 OZ 190 0.00 8 TG 76 13.16 JQ 80 30.00 9C 119 10.92 QF 179 4.47 9 VN 76 18.42 MH 80 5.00 MI 98 18.37 KE 168 0.00 10 CX 71 28.17 MI 76 13.16 MF 92 32.61 VN 164 3.66 11 JQ 62 32.26 TG 74 5.41 KA 81 0.00 MF 161 14.91 12 MI 62 16.13 CX 69 26.09 TG 80 7.50 VA 159 3.77 13 JL 56 3.57 KA 66 0.00 NH 74 0.00 CX 148 6.76 14 NH 56 0.00 NH 64 0.00 FD 72 0.00 MI 141 9.93 15 NZ 52 1.92 5 J 62 0.00 5 J 70 0.00 EY 134 1.49 16 QF 52 11.54 JL 52 0.00 PR 70 5.71 AK 132 0.00 17 SQ 50 0.00 SQ 52 0.00 MH 68 2.94 9C 129 18.60 18 GA 48 16.67 QF 48 8.33 JQ 66 12.12 GA 124 3.23 19 PR 45 13.33 GA 46 8.70 CX 64 21.88 HU 120 6.67 20 KA 42 0.00 NX 46 0.00 TR 62 0.00 TG 117 3.42
The results of hub identification and layout evolution of the typical carriers' RCEP international airline networks are obtained based on the proposed method. It is identified that the core hubs of MU in the RCEP international airline network in 2010 were PVG, HKG, and KMG, which are all located in China and overlap with the regional airline network's two hubs. Their own contributions proportion are shown in Table 7. With the expansion of network scale, the core hub role of PVG is becoming increasingly significant, but the hub role of airport HKG and KMG is gradually decreasing. The overall network is centered around airport PVG, with airports HKG, KMG, and other regional hubs presenting a 'one core, multiple zones' layout structure distribution, gradually becoming more similar to the structural distribution of the regional airline network.
Table 7 The core-hub airports of the RCEP international airline network of typical carriers from 2013 to 2019.
Airlines Airport name 2010 (%) 2013 (%) 2016 (%) 2019 (%) MU PVG (Shanghai Pudong International Airport) 61.58 59.86 73.86 81.42 HKG (Hong Kong International Airport) 5.11 3.48 1.34 0.48 KMG (Kunming Changshui International Airport) 1.27 2.62 3.65 2.20 OZ ICN (Incheon International Airport) 99.25 99.35 99.46 98.63 PUS (Gimhae International Airport) 49.81 59.11 42.60 80.11 AK KUL (Kuala Lumpur International Airport) 93.06 94.97 96.78 96.27 BKI (Kota Kinabalu International Airport) 2.09 2.75 16.69 23.32 9C PVG (Shanghai Pudong International Airport) 58.18 72.83 BKK (Suvarnabhumi Airport) 7.26 8.32 KIX (Kansai International Airport) 4.38 2.39 YTY (Yangtai International Airport) 0.27 18.39
From a topological analysis, the RCEP international airline network of OZ has core hub airports, including ICN and PUS, both located within South Korea and overlap with the regional airline network's one hub. According to Table 7, the core hub role of airport ICN has remained significant in the past decade, while the size of airport PUS gradually increasing. The "dual-hub" layout of OZ in the RCEP international airline network is becoming more stable.
Similar to OZ, AK has core hub airports KUL and BKI in the RCEP international airline network, which are located within Malaysia, overlapping with the regional airline network's one hub. In the past decade, the hub role of airport KUL has become increasingly stable, while airport BKI has gradually become a core hub airport since 2016 with the investment and construction of AK. The development of the RCEP regional economy and the expansion of airline markets have had a positive impact on each other.
Spring Airlines has many core hub airports, scattered throughout China, Japan, and Thailand. The core hub airport is mainly PVG, supplemented by BKK, KIX, and YTY, overlapping with the regional airline network's two hubs. The hub role of airport BKK is slowly increasing, with a decrease in airport KIX and a significant increase in airport YTY, indicating that Spring Airlines intends to transfer its core hub from abroad to China.
Comparing the layouts and evolution of typical carriers' RCEP international airline networks, it is found that carriers often choose local airports in RCEP international aviation hubs as core hubs or base airports to develop and operate their airline networks, which is caused by political and traffic rights factors. In addition, the development of RCEP regional economies will also attract airlines to build and develop regional hub airports domestically, which is conducive to promoting the development of regional airline networks. The carriers' network and regional network have a positive impact on each other.
Figure 5 shows the relationship between airport importance and node activity in the RCEP international airline network over the past 10 years. Node activity positively correlated with airport importance, and airports of high importance tend to aggregate. This is mostly due to the influence of factors such as the economic situation and government-determined traffic rights. Carriers tend to choose large international airports in the political and economic centers of countries, which increases the importance of carriers operating flights in regional international aviation hubs. Besides, some airports have low airport importance but high node activity, such as Perth International Airport and Adelaide International Airport in Australia. This is mostly because carriers prefer to develop and operate routes around their base airports, which results in many local carriers competing for the transportation market of the few routes at the base airport. Last but not least, some airports are selected as bases or hubs only by low-cost carriers because full-service and low-cost carriers have different preferences in airline network layouts. There are therefore several airports exhibiting high airport importance but low node activity.
Graph: Figure 5 Relationship between airport importance and node activity from 2013 to 2019.
To compare the similarities and differences in the route layouts of carriers within the same airport node over the past 10 years, Table 8 lists the airport node types of the RCEP international airline network divided according to the multiplex participation coefficient. It is observed that the number and proportion of centralized and mixed airport nodes are gradually decreasing, while the number of multi-layered airport nodes has doubled in the past 10 years, and by 2019 accounted for 71.6% of the total number of airport nodes in the RCEP international airline network. This suggests that the continuous expansion of the RCEP international airline network has produced an evolutionary trend of airport nodes developing into the multi-layered type and most airports are in similar operating positions and provide similar transit services for passengers in routes used by different carriers. It can be inferred that carriers would follow the existing route layout structure in the RCEP international air transportation market when selecting airports or launching new routes, thus having similar international route layout structures. Additionally, 19.9% of airports in 2019 were centralized nodes, indicating that there are also many airports with routes monopolized by one or only a few carriers in the RCEP international airline network.
Table 8 Airport node types in the RCEP international airline network from 2013 to 2019.
Type 2010 2013 2016 2019 Number Proportion (%) Number Proportion (%) Number Proportion (%) Number Proportion (%) Centralized 52 32.7 44 26.3 41 21.4 42 19.9 Mixed 32 20.1 36 21.6 36 18.8 18 8.5 Multi-layered 75 47.2 87 52.1 115 59.9 151 71.6
Figure 6 shows the association between the multiplex participation coefficient and the importance of airports in the RCEP international airline network. The core-hub, key-spoke, and local-hub airports whose airport importance are greater than 1 are all considered multi-layered nodes, while the local-hub airports and feeder-line airports with airport importance of less than 1 could be considered mixed or centralized nodes. The findings suggest that the whole RCEP international air transportation market facilitates fair and open competition and a cooperative environment. However, the market and route competition is more fierce among hub airports of large importance because they usually attract more carriers. While feeder-line airports and new routes in the network have low influence, their development may increase in the future. Moreover, the number of airport nodes with high importance but zero multiplex participation coefficient has been decreasing in the past decade. In 2019, the multiplex participation coefficient of airport nodes with node importance greater than 1 was highly concentrated, indicating that the positioning and division of labor of each airport in the airline network are becoming increasingly clearer.
Graph: Figure 6 Relationship between airport importance and multiplex participation coefficient from 2013 to 2019.
The types, full names, and abbreviations of typical carriers are shown in Table 9. The relationship between the route uniqueness of typical airlines in the RCEP international airline network in the last decade and its number of routes is shown in Fig. 7. It is found that the route uniqueness of the airlines with the higher number of routes tends to be close to 0.5, indicating that the large airlines have about half of the routes independently operated and half of the routes competing in the RCEP international airline network. While most airlines have been experiencing steady growth in the number of route edges, several large full-service carriers had significantly lower route uniqueness in 2019.
Table 9 The types, full names, and abbreviations of typical carriers.
Types Carrier Carrier Code Types Carrier Carrier Code Full-service carrier China Eastern Airlines MU Full-service carrier Korean Air KE Full-service carrier China Southern Airlines CZ Low-cost carrier Spring Airlines 9C Full-service carrier Air China CA Low-cost carrier Air Asia AK Full-service carrier Asian Airlines OZ
Graph: Figure 7 Relationship between route number and route uniqueness of typical carriers from 2013 to 2019.
The relationship between the route uniqueness of typical airlines in the RCEP international airline network in the last decade and its number of routes is shown in Fig. 7. It is found that the route uniqueness of the airlines with the higher number of routes tends to be close to 0.5, indicating that the large airlines have about half of the routes independently operated and half of the routes competing in the RCEP international airline network. While most airlines have been experiencing steady growth in the number of route edges, several large full-service carriers had significantly lower route uniqueness in 2019.
The number of routes of full-service airlines MU, CZ, CA, OZ, and KE rank high and show an overall trend of growth, but the route uniqueness gradually decreases with the increase of the number of routes. This suggests that the new international routes in the RCEP region added by large full-service airlines after 2016 are mostly existing routes in the transportation market, and there are fewer new routes. Obviously, full-service airlines increased the frequency of flights on existing international routes, improved transportation efficiency in the RCEP international airline network, and gained benefits through transportation density economies.
Low-cost carriers have also strengthened their international airline networks in the RCEP region rapidly compared to full-service carriers. AK's routes have increased 1.5 times over 10 years. 9C's routes have increased 16 times from 2013 to 2019. The route uniqueness of AK and 9C has been maintained at 0.5 or below, indicating that low-cost carriers are more attentive to opening new routes than full-service carriers, and are adept at utilizing the scale effect economy to gain greater benefits by monopolizing the market of certain international routes in the RCEP region, and are the main force in the continuous expansion of the RCEP international airline network.
Building an international aviation corridor will be an important measure to improve national competitiveness after the implementation of the RCEP agreement. It can be foreseen that competition among international aviation hubs in the RCEP region will also become more intense. At present, the hubs of the RCEP international airline network are distributed in Singapore, South Korea, Malaysia, Thailand, and China.
Firstly, stable aviation hubs such as Singapore Changi Airport, Kuala Lumpur International Airport, Incheon International Airport, and Hong Kong International Airport are recommended to maintain their central hub position and highlight their competitiveness in the international area. Secondly, Beijing Capital International Airport, Kansai International Airport, and Guangzhou Baiyun International Airport are international aviation hub preparations for the RCEP region and are important bases for carriers to build airline networks for the RCEP international market. Therefore, it is necessary to seize the opportunities of the implementation of the RCEP agreement to improve internationalization. Finally, based on hub identification, it is found that the main base carriers are often the biggest supporters and beneficiaries of hub construction. Therefore, it is necessary for airports to cooperate with carriers in the development of hub construction, maximize cooperation efficiency, help carriers expand their airline networks, and improve the convenience of transferring passengers and cargo.
Although the RCEP agreement has stimulated and expanded the demand and market for international air transportation, the number of carriers participating in the RCEP international air transportation market is also increasing, and competition in the top market will become more intense with the expansion of the RCEP international airline network. This may also bring changes to the market structure and layout for the development of air transportation. Therefore, carriers' airline network layout planning must be more cautious.
According to the analysis of the size and evolution of the RCEP international airline network, it is found that the development of routes is mainly based on direct routes. However, the air transportation market is vast and the demand for outbound tourism is high. The development of stopover routes is not only beneficial for carriers to concentrate domestic transportation demand and form a certain scale economy effect but also for promoting the transformation of domestic transportation demand to international transportation demand, promoting a new development pattern of mutual promotion between domestic and international dual circulation in the air transportation industry, effectively improving circulation efficiency and reducing circulation costs. At the same time, based on the analysis of the multi-layered characteristics, most carriers have similar RCEP international airline layouts. When choosing transportation airports or adding new routes, they will pay attention to referring to existing routes. However, the competition for existing routes is intense, especially in aviation hubs and some base airports. In this case, considering stopover routes is beneficial for improving flight occupancy rates, and increasing the economic benefits and market competitiveness of carriers. Therefore, in the early stages of carriers' airline network development, carriers can build a certain proportion of stopover routes around base airports if hub competition is fierce.
Secondly, by analyzing the uniqueness of routes, it is found that low-cost carriers are more focused on opening up new RCEP routes than full-service carriers. The RCEP international airline network monopoly of large full-service carriers has reduced and increased the frequency of flights on existing international routes since 2016. This reflects the two strategies of carriers to expand their routes. One is to rely on opening new international routes in the RCEP region around the base airport and monopolizing this market, enhancing their market share in the regional international air transportation market. Another approach is to consolidate existing routes and increase flight frequency to achieve economic benefits. The former is suitable for new carriers and low-cost carriers entering the RCEP international air transportation market, while the latter is more suitable for large full-service carriers. However, while carriers continue to expand their routes and improve their hub layout, they also need to focus on building more reasonable flight schedules, fully leveraging the transfer and convergence role of hub structures, and improving the level of passenger and cargo transportation services.
To analyze the layout of the RCEP international airline network and provide recommendations for the development of hub airports, full-service carriers, and low-cost carriers in the RCEP region, this paper identifies the regional aviation hub of the RCEP international airline network with the proposition of a method, which is based on an improved contribution matrix and comprehensively considers the contribution of neighboring airports and the multi-layered characteristics of the airports. The rationality of the proposed method is comparatively verified by the Susceptible-Infectious-Recovered (SIR) model. The proposed method is then used to identify the hub airports of the RCEP international airline network, and further analyze the layout of the RCEP international airline networks from overall and carriers' perspectives and the multi-layered characteristics of airports and routes. Suggestions for the development of hub airports and the layout planning of full-service carriers and low-cost carriers in the RCEP international airline network are provided ultimately.
The following main results are obtained:
This study was funded by the National Natural Science Foundation of China, Grant Number 52372298.
Conceptualization, W.Y.; Data curation, Y.C. and Y.H.; Formal analysis, Y.C. and Y.H.; Funding acquisition, W.Y.; Methodology, W.Y., Y.C., Y.H., W.W. and Z.X.; Project administration, W.Y. and W.W.; Resources, W.Y.; Software, Y.C.; Supervision, W.Y., W.W. and Z.X.; Visualization, Y.C. and Y.H.; Writing–original draft, Y.C. and Y.H.; Writing–review & editing, Y.C., W.Y. and W.W.. All authors reviewed the manuscript.
The datasets generated during and/or analyzed during the current study are available from the Official Airline Guide (https://oag.cn/).
The authors declare no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
By Wendong Yang; Yulin Chi; Yining Huang; Wenbin Wei and Zhengjia Xu
Reported by Author; Author; Author; Author; Author