Endogenous control of service rates in stochastic and dynamic queuing models of airport congestion

Airport congestion mitigation requires reliable delay estimates. This paper presents an integrated model of airport congestion that combines a tactical model of capacity utilization into a strategic queuing model. The model quantifies the relationships between flight schedules, airport capacity and flight delays, while accounting for the way arrival and departure service rates can be controlled over the day to maximize operating efficiency. We show that the model estimates the average and variability of the delays observed at New York’s airports relatively well. Results suggest that delays can be extremely sensitive to even small changes in flight schedules or airport capacity.     

Determinants of flight delays at East Asian airports from an airport,route and network perspective

Previous studies on determinants of flight delays have been limited to the attributes of departure and arrival airports and routes and experiences in North America and Europe. This research extends the existing discussions by considering a network perspective to explore East Asian airports. A total of 4611 routes among 318 airports in the second half of 2017 are analyzed using panel data regressions. Results show that the attributes of airports and routes linked with the departure and arrival airports of a flight are the essential determinants of flight delays in East Asia. However, the congestion internalization and hubness effect of departure and arrival airports suggested in previous research are not observed in this region. The associations of slot control levels and route attributes with delays are varying between airports inside and outside China and domestic and international routes. This research supplements a new perspective to understand the determinants of flight delays in existing literature and provides a rarely observed knowledge to encourage East Asian airport operators and carriers developing management strategies for minimizing flight delays.     

Effect of airline choice and temporality on flight delays

The civil aviation industry is critical to the economic development of any nation. Being a key infrastructure service, the performance of the civil aviation industry spills over to other segments of the economy. We investigate service failure among US airlines, using departure delay, and model its likelihood using mixed-logit regression. We characterize arrival performance for delayed departures using a service recovery/double deviation framework. We model the likelihood of different arrival possibilities using multinomial logistic regression. We use the chosen airline, temporality, and flight duration as predictors. Our key results include: (i) Afternoon/evening flights, and longer duration flights are more likely to depart late; (ii) Delta Airlines is most likely to depart on time; (iii) Southwest Airlines is best in recovering from delayed departures; (iv) United Airlines is most likely to aggravate delays.     

An Exploratory analysis of flight delay propagation in China

Delay propagation is the flight departure delay caused by the arrival delay of pre-segment flight. Chinese airline market has suffered very poor on-time performance (OTP) in recent years. It is, however, unclear whether delay propagation prevails as one major source for such problem. This study first aims to empirically quantify delay propagation in the Chinese airline market. Specifically, we shed light on heterogenous levels of delay propagations across different airports and airlines. Then, the distinct delay propagation patterns in China are also discussed and compared with other developed airline markets (e.g., the US and Europe). Our estimation is based on OTP data for over 12 million Chinese flights covering the 2015–2017 period. Specifically, it is found that 10 min arrival delay of pre-segment flight within 1 hr before the departure lead to an average of 7.49 mins delay propagation for subsequent departure flight. Arrival delay of earlier pre-segments (1–2 and 2–3 hr before the departure) leads to much less delay propagation, due to longer ground buffer. Chinese airlines arrange longer ground and flight buffer than that of the US airlines to prevent the delay propagation from accumulating along the subsequent flights in a day. Thus, unlike the US market, delay propagation is not the major reason for poor OTP in China. In addition, delay propagation is less prevailing at the Chinese hub airport. This is because China has relied on point-to-point network, which does not require sophisticated schedule coordination. And the local passengers at these Chinese hub airports have higher time value such that the Chinese airlines also try to improve OTP at these hub airports to better serve these lucrative but time-sensitive local passengers. Unlike the European LCCs, Spring Airlines, the largest low-cost carrier (LCC) in China, outperforms major full-service carriers (FSCs) in controlling delay propagation. This finding may also apply to other Northeast Asian LCCs sharing common operational characteristics as Spring Airlines. Last, we find that airlines purposely tolerate moderate departure delays of up to 15 min, which is the threshold that defines delays, no matter whether the pre-segment flight arrives late or on-time. The relevant policy and managerial implications are also discussed.     

Airport and route classification by modelling flight delay propagation

Flight delay can be divided into the root and propagated delays for identification and analysis of airports/routes that have a great influence on actual flight delay using flight operations raw data (tower logs) for South Korean domestic flights. In addition, the presented concept of “generated delay” classifies airports as affecting other airport delays or affected by other airport delays. The generated delay refers to the delay time at a particular airport/route, and this delay thereafter actually propagates to other airports/routes. In this study, the generated and propagated delays were displayed on a two-dimensional graph, and the airports/routes were grouped according to delay characteristics. Group A represented airports with high generated and propagated delays, and group B represented airports that are highly affected by delays of preceding airports/routes. Group C represents airports where newly formed delays affect other airport delays, but these airports mitigate delays from other airports. Airports in group D have relatively low delays and propagation. Thus, we targeted airports belonging to groups A and C because of their delay propagation impact on other routes/airports, which must be reduced by decreasing the root delay from targeted routes/airports. Among the airports, Jeju international airport (CJU) had the highest average delay time and propagated flight delay time with similar averaged generated delay times. Among the routes, departure flights from various airports to CJU had significant propagation effects on the subsequent flights. CJU and related routes have a very large impact on domestic flight delays because South Korean domestic airline routes are concentrated on CJU. However, there has been no quantitative analysis, and it is meaningful that the quantitative analysis results were presented in this study. In addition, we suggest that other airports such as GMP (Gimpo), CJJ (Cheongju), WJU (Wonju), and KUV (Gunsan) have a significant impact on domestic flight delays.     

Using causal machine learning for predicting the risk of flight delays in air transportation

Delays in air transportation are a major concern that has negative impacts on the airline industry and the economy. Given the complexity of the National Air Space system, predicting the risk of flight delays and identifying significant predictors is vital to risk mitigation. The purpose of this paper is to perform data mining using causal machine learning algorithms in the USELEI process (Understanding, Sampling, Exploring, Learning, Evaluating, and Inferring) to predict the probability of flight delays in air transportation using data collected from different sources. The findings indicated significant effects of predictors, including reported arrivals and departures, arrival and departure demands, capacity, efficiency, and traffic volume at the origin and destination airports on the risk of flight delays. More importantly, causal interrelationships among variables in a fully structural network are presented to how these predictors interact with one another and how these interactions lead to delay incidents. Finally, sensitivity analysis and causal inference can be performed to evaluate various what-if scenarios and form effective strategies to mitigate the risk of delays.     

A model for projecting flight delays during irregular operation conditions

On-time performance of airlines schedule is key factor in maintaining current customer satisfaction and attracting new ones. This requires management of the different operation resources (crew/aircraft) to ensure their on-time readiness for each flight in the planned schedule. However, flight schedules are often subjected to irregularity. In particular, weather accounts for nearly 75% of system delays. Due to the tight connection among airlines resources, these delays could dramatically propagate over time and space unless the proper recovery actions are taken. This paper presents a model which projects flight delays and alerts for possible future breaks during irregular operation conditions. The results of applying the model at the operation control center of a major airlines company in the US are presented.     

Has market concentration fostered on-time performance? A case study of seventy-two U.S. airports

The study compares a multivariate with a quantile regression model to determine whether utilized airport capacity, departure and airborne delays, departure and arrival demand, and market structure explained variations in on-time gate arrivals and arrival delays. In both models, airport departure delays, arrival and departure demand explained variations in the two response variables in prioritized and non-prioritized metroplexes, holding other variables constant. After 2008, the consolidation of the airline industry through mergers coincided with the implementation of NextGen programs, which may have contributed to improvements in on-time performance, especially at prioritized metroplexes where airspace was redesigned.     

Network-wide assessment of ATM mechanisms using an agent-based model

This paper presents results from the SESAR ER3 Domino project. Three mechanisms are assessed at the ECAC-wide level: 4D trajectory adjustments (a combination of actively waiting for connecting passengers and dynamic cost indexing), flight prioritisation (enabling ATFM slot swapping at arrival regulations), and flight arrival coordination (where flights are sequenced in extended arrival managers based on an advanced cost-driven optimisation). Classical and new metrics, designed to capture network effects, are used to analyse the results of a micro-level agent-based model. A scenario with congestion at three hubs is used to assess the 4D trajectory adjustment and the flight prioritisation mechanisms. Two different scopes for the extended arrival manager are modelled to analyse the impact of the flight arrival coordination mechanism. Results show that the 4D trajectory adjustments mechanism succeeds in reducing costs and delays for connecting passengers. A trade-off between the interests of the airlines in reducing costs and those of non-connecting passengers emerges, although passengers benefit overall from the mechanism. Flight prioritisation is found to have no significant effects at the network level, as it is applied to a small number of flights. Advanced flight arrival coordination, as implemented, increases delays and costs in the system. The arrival manager optimises the arrival sequence of all flights within its scope but does not consider flight uncertainties, thus leading to sub-optimal actions.     

The impact of flight delays on passenger demand and societal welfare

US airline passengers increasingly have access to flight delay information from online sources. As a result, air passenger travel decisions can be expected to be influenced by delay information. In addition, delays affect airline operations, resulting in increased block times on routes and, in general, higher carrier costs and airfares. This paper examines the impact of flight delays on both passenger demand and airfares. Delays are calculated against scheduled block times as well as against more idealized feasible flight times. Based on econometric estimations, welfare impacts of flight delays are calculated. We find that flight delays on a route reduce passenger demand and raise airfares, producing significant decreases in both consumer and producer welfare. Since producer welfare effects are estimated to be three times as large as consumer welfare effects, we conclude that from an economic efficiency rationale, airlines should be required to pay for the bulk of flight delay remediation efforts.     

Analysis of airport weather impact on on-time performance of arrival flights for the Brazilian domestic air transportation system

Air traffic operations are significantly impacted by weather conditions. These external factors may impose operational constraints and generate demand-capacity imbalances, leading to reduced on-time performance, additional airline costs and inconveniences to passengers. Efficient management of such disruptions requires an understanding of the main causes of flight delays towards increasing their predictability. In this study, we investigate the impacts of airport surface weather conditions on the likelihood of flight delays for the Brazilian domestic air transportation system. We use historical flight schedule, on-time performance and weather data and estimate a logit model to analyze how different meteorological variables at the airport of destination affect the probability of a delayed arrival. We obtain empirical evidence for the impacts of low ceiling and visibility conditions, precipitation and wind gusts on the likelihood of arrival delays for the set of Brazilian airports analyzed.     

An analysis of Brazilian flight delays based on frequent patterns

In this paper we applied data indexing techniques combined with association rules to unveil hidden patterns of flight delays. Considering Brazilian flight data and guided by six research questions related to causes, moments, differences, and relationships between airports and airlines, we evaluated and quantified all attributes that may lead to delays, showing not only the main patterns, but also their chances of occurrence in the entire network, in each airport and airline. We observed that Brazilian flight system has difficulties to recover from previous delays and when operating under adverse meteorological conditions, delays occurrences may increase up to 216%.     

Dynamic routing of time-sensitive air cargo using real-time information

The route planning of time-sensitive air-cargo is becoming more important with the growing air-network congestion and delays. We consider a freight forwarder’s routing of a time-sensitive air-cargo in the presence of real-time and historical information regarding flight availability, departure delays and travel times. A departure delay estimation model is developed to account for real-time information inaccuracy. A novel Markov decision model is formulated and solved with online backward induction. Through synthetic experiments and case studies, we demonstrate that dynamic routing with real-time information can improve delivery reliability and reduce expected cost.     

A tentative analysis of the impacts of an airport closure

This paper identifies the vulnerabilities of operations at Heathrow airport to a short airport closure. The system disruptions assessed are delays, flight rerouting or diversions to alternate airports, and flight cancellations. In addition, the impact on fuel consumption, and hence CO₂ emissions, is quantified using the Reorganized Air Traffic Control Mathematical Simulator Plus simulation model and the Advanced Emission Model tool. The results have implications for assessing how increased severe weather events associated with climate change may disrupt airport operations.     

Analyzing tactical control strategies for aircraft arrivals at an airport using a queuing model

This paper proposes to analyze control strategies for arrival air traffic at an airport using a classical queuing model. The parameters of our model are estimated by means of a data-driven analysis of two years of radar tracks and flight plans for arrival flights at Tokyo International Airport from 2016 to 2017. Our results show that increasing the capacity with one or two more aircraft in the airspace up to 60 NM around the airport significantly mitigates arrival delays, even when assuming future, increased arrival traffic volumes. The outcomes of this study provide insights into the effectiveness of arrival control strategies and are seen as a means to recommend scenarios to be further analyzed with human-in-the-loop simulations.     

A rolling horizon approach to the high speed train rescheduling problem in case of a partial segment blockage

This paper reschedules train services on a double-track high speed railway in a disrupted situation, where one track of a segment is temporarily unavailable. We have to decide the sequence of train services passing through the blocked segment, the arrival and departure time of each train service at each station, and the canceled train services. Three practical train rescheduling strategies are explicitly compared and formulated by three MILP models. The uncertain duration of the disruption is handled. A rolling horizon approach is applied to solve our models. The models are tested on a real-world instance of the Beijing-Shanghai high speed railway.     

Clustered airline flight scheduling: Evidence from airline deregulation in Korea

This paper explores the impacts of competition level on airline scheduling in the Korean domestic short-haul routes where a hub-and-spoke system is not the optimal air transport network strategy. The empirical findings using the Korean airline panel data for the period 2006–2010 suggest that competition leads to less differentiated departure flight times as expected from spatial competition theory. Unlike the previous study on the U.S airline industry, the degree of this tendency for less differentiation differs across the type of routes: the Jeju island routes (leisure type) and the inland routes (business type), in the deregulated period. Following the May 2008 Deregulation Act we find an increasingly clustered pattern of airline scheduling in the Jeju island routes where there have been competitive pressures associated with new low cost entrants. This recent evidence would imply that airline carriers strategically schedule departure flight times and allocate flights between routes as competition increases in the deregulated period.     

Modeling the flight departure delay using survival analysis in South Korea

The air transportation industry in South Korea has been growing since the 2000s. Although the number of air passengers has continuously increased, the flight delay rate has fluctuated from year to year. In this study, a survival analysis was conducted to compare each flight delay in terms of airport/time slot and to evaluate the actual impacts of major variables on flight delays. We performed multiple analyses using survival analysis methodologies, including the Kaplan–Meier estimator, log-rank test, and Cox proportional hazards model. In particular, we applied a turnaround buffer as a key variable in Cox proportional hazards model analysis. The survival curve showed that more international flights departed earlier than scheduled or with minimal delay than domestic flights. However, international flights tended to have more long-term delays. The log-rank test indicated that international flights had a wider distribution of delay times than domestic flights. In the Cox proportional hazards analysis, under the condition in which the actual flight turnaround time was less than 120 min, the results indicated that the longer the actual turnaround buffer was, the less frequently the delays occurred. By using these analyses, we could numerically verify the actual delay trends in South Korea. The results can be used as fundamental resources for political and economic decision-making processes in the aviation industry.     

Modelling airline flight cancellation decisions

In order to predict airline responses to Traffic Management Initiatives (TMIs), and reveal the underlying preference structures that shape these responses, we study US domestic airlines’ cancellation decisions in response to the Federal Aviation Administration (FAA)’s TMIs, in particular, to Ground Delay Programs (GDPs). By observing the actual flight-cancellation choices made by airline dispatchers, the airlines’ cancellation utility functions can be inferred through the use of binary choice models. The model captures how delays to a given flight and potential delay savings to other flights affect flight cancellation decisions. We also find larger, fuller, less frequent, shorter-distance, and spoke-bound flights are less likely to be cancelled, and that there is inter-airline variation in flight cancellation behaviour.     

Flight delay impact on airfare and flight frequency: A comprehensive assessment

This paper presents a comprehensive empirical analysis of flight delay impact on airfare and flight frequency in the US air transportation system. We model airfare and flight frequency as functions of cost and demand characteristics, competition effects, and flight delays at origin, destination, and intermediate hub airports. Estimation results confirm that airlines tend to pass delay cost onto passengers through higher fare, whereas delay has an upward effect on flight frequency. We find that proportionate airport delay reduction across the system can result in annual fare reduction in the order of billion dollars.