Technical Program Abstracts
Airline On-Time Performance (OTP) Analysis
On-Time Performance (OTP) is not only an operations and marketing relevant KPI for airlines, it is also a good economics performance indicator that can depict their effectiveness and efficiency gains. Flights are a central element of the production function of an airline. When the amount of monthly flights for an airline changes, the OTP changes as well, but often in different direction and with different intensity. We establish the interpretation of changing OTP relative to the change of the output level of the production function. Then, we continue by presenting the global cost estimation study conducted for over 130 global airlines. We introduce a cost estimation model, to complement the understanding of OTP competitiveness, and to present the costs of delays incurred by these airlines. We conclude by discussing features of delay management systems for improved OTP competitiveness.
An integrated optimization approach to the challenge of aircraft maintenance planning
Luis Alvarez - INFORM
As more and more aircraft take to the skies and engineering skills remain in short supply, the future of aircraft maintenance will require innovative ideas to continue carrying out maintenance efficiently and safely. Tail assignment, maintenance staff planning and work package planning play central roles in how overall maintenance is carried out. Although each of these depend on each other, optimization solutions have so far looked at them only individually. Advances in hardware and optimization algorithms open up new opportunities to start taking an integrated optimization angle involving all three areas and offering benefits not possible until now. In this talk we will explore the challenges associated with taking an integrated view, discuss the possible solutions and present some first results showing the promises of this approach.
24/7 Optimisation of Operational Slot Assignment
Valentin Weber - Amadeus
To manage airport congestion, some ANSPs (Air Navigation Service Providers) use GDPs (Ground Delay Programs) in their major airports. Airport capacity is represented by time slots and airlines can exchange slots through a CDM (Collaborative Decision Making) system to make their flights compliant to depart. Slot management is today mostly done manually by dedicated operators. To help them, we developed a decision support tool based on continuous optimisation that automatically reassigns slots in order to minimise globally the ATFM (Air Traffic Flow Management) delay due to slots for the airline. We will present how we successfully deployed this live application inside an Operations Control Center and what are now the benefits for the airline.
A Discrete Event Simulation Model for Airport Capacity Planning
Douglas Smith - University of Missouri-St. Louis
We present a discrete-event simulation model built in Arena 14.7 for airside activity at commercial airports. The model helps airport planners (and potentially airline operations analysts) to assess the impact of changes in airport facilities (runways, taxiways, ramps, gates); aircraft activity (scheduled and unscheduled); resources deployed (ground equipment and services); and operating practices (scheduling and dispatching strategies by airline operations, ground control and ATC). The model represents airport activity as a system of staged queues. It includes aircraft movements from final approach fix to touchdown, in the network of taxiways and on ramps; gate and ramp activity; and departures. We discuss model calibration, detailed performance metrics generated for individual carriers, and applications to several operating scenarios (including local weather delays and traffic delays due to congestion at connecting hubs).
Airline Maintenance Strategies:“ In-House vs. Outsourced“ An Optimization Approach
Massoud Bazargan - Embry-Riddle Aeronautical University
Currently, on average, aircraft maintenance accounts for more than 10% of total operating and 17% of total aircraft operating costs.
This study offers a new mathematical modeling approach to help airlines identify which types of heavy aircraft maintenance checks be performed in-house or outsourced. This study attempts to minimize the total cost of heavy maintenance programs over a planning period subject to performing all maintenance programs on time and other side constraints. The model is applied to six airlines, three US and three European with different network sizes and business models. The results are very encouraging and somewhat counter-intuitive. The solutions recommend that more expensive and labor intensive checks be outsourced. The cost breakdown indicates that in the long run, fix cost capital investments for in-house maintenance are dominated by cheaper outsource costs. "
Commercial Steering of Operations Control
Matthias Feix - m2P Consulting
The Operations Control Center is the heart of each airline. Working 24/7, 365 days a year, the goal is to respond to irregularities of all kinds to keep the schedule intact and maximize on-time performance. To achieve this goal hundreds of decisions are taken every day which have a direct impact on aircraft, crews, customers, revenues and cost of an airline. Currently decisions are mostly taken based on the experience of the individual controller who focuses on operational stability with limited transparency about the potential financial impact. With airlines already leveraging sophisticated systems to optimize their revenues, the question is whether data and analytics driven approach could be introduced to help controllers take more profitable decisions. The presentation will elaborate the concept of commercial steering in operations control, requirements, success factors and limits of the concept and an open discussion of chances based on specific use cases.
Employee Insights - Finding the Unsung Heroes in our Frontline Workforce
The most granular level of reporting for operational metrics has generally been at the airport level. To give airports more visibility, American Airlines has created scorecards that look at individual employee performance. In this presentation, we will discuss how American Airlines’ hubs and gateways are using dashboards for Gate Agents and Ramp Leads to quantify employee performance and improve dependability.
From Twitter to Airline Operations - An Exploratory Study on Social Media Data Mining
Every day thousands of airline customers express their opinions through Twitter. This customer feedback from Twitter is a new source of information, and there is a potential to see if analyzing these messages can support airlines' operational decisions. The Advanced Analytics Group at Boeing Canada sought to explore this potential through a partnership with the Centre for Operations Excellence (COE). The team verified that Twitter sentiment is a reliable indication of customer satisfaction with airlines. Moreover, through a statistical model that predicts negative sentiment on Twitter based on flight disruption attributes, the team demonstrated how Twitter data can be used to support disruption management decisions. The model helps airlines prioritize among different flights to minimize disruption complaints on social media.
Hybrid Particle Swarm Optimization with Parameter Fixing – Application to Automatic Taxi Management
Giuseppe Sirigu - Georgia Tech
The conflicting needs of meeting the growth of air traffic, reducing the additional air pollution and costs deriving from congestion at airports, push towards the modernization of ground operations and airport management. In this framework, a new solution is proposed to perform just in time taxi operations using autonomous electric towbarless tractors. The purpose of this solution is to eliminate queues and to reduce the environmental and economic impact of ground operations, meeting the requirements for the future air traffic management (ATM). A hybrid particle swarm optimization (HPSO) algorithm was developed to provide conflict-free schedules for the tractor autopilots. In order to improve the rate of convergence of the algorithm, we developed a parameter fixing algorithm, which constrains the particle elements, based on the particle history.
Leveraging Predictive Analytics in Airline Operations Control
Alex Huang - The Weather Company
Airport delays are often avoidable and the industry has seen how poor decision making can lead to operational issues that ultimately impact the bottom line. One common issue for airlines is integrating big data into their operations for real-time decision making. In this session, the speaker will examine factors that directly and indirectly affect airport operations and address the potential impact of delays at airports to airlines. He will discuss how airlines can combine data science, analytics and weather to improve cost savings by streamlining airport operations, and present modeling approaches for three key airport performance metrics: runway configurations, airport congestion and taxi time. Attendees will leave with a better understanding of how pairing historical airport operational data with superior weather forecasting can lead to actionable predictions and how predictive analytics can be applied to daily operations to optimize critical decisions.
Lower Cost Airport Departure Operations under the Departure Metering Concept
Heng Chen - University of Massachusetts Amherst
Departure metering is an airport surface management procedure that limits the number of aircraft at the runway queue by holding aircraft at gates or at a predesigned metering area. In this paper we propose a stochastic dynamic programming framework to identify the optimal gate, metering area and departure queue allocation policies to minimize expected overall fuel burn costs. We also provide specific policies given different airport configurations. We further identify the optimal capacity for such a metering area, and quantify the overall value of the presence of a departure metering area at airports. Potential fuel savings are calculated for representative airports and airlines.
Modeling Payload Weight Variance for Cruise Optimization
Charles E. Kaul - Boeing Commercial Aviation Services
Bayesian Statistical models derived from repeated sampling of in-service Airline QAR data are employed to develop tail-specific parametric models of cruise flight. The models provide 'implied' performance weight estimates to improve recommended Econ Cruise speeds and Optimal Altitudes. These estimates are conditioned on overall tail-specific performance metrics vs. published OEM values, overall route-specific factors and flight-specific payload weight bias. Significant fuel consumption and carbon emissions efficiency gains are expected.
Operational Cost Model: Enhancing Quality of Complex Decision Making
Dong Liang - Sabre Airline Solutions
The primary goal of an Airline Operations Center is to operate the airline's flight network efficiently and safely. This task involves decision makings. In this talk, we introduce the new Operational Cost Model (OCM), which provides an innovative approach to forecast, calculate, share, and report an airline's operational disruption costs in order to improve decision-making quality during daily operations. OCM gives airlines potential for substantive performance and operational cost management. We also propose the concept of Operational Cost Performance (OCP), complementary to On-Time Performance (OTP), to evaluate the success of airlines' daily operations.
The Baker: A Recovery Optimization Engine at Southwest Airlines
Charles Cunningham/Ryan Files - Southwest Airlines
Uses and impacts the of the optimization engine â€œThe Bakerâ€. This revolutionary tool converted the best practices of the group charged with handling the airlines daily IROPs and recovery problems, into a complex algorithm that is used to solve many types of operational disruptions. The Baker solves station shut down or reduction problems encouraging proactivity. When weather dictates the need to suspend operations, proactivity has had a dramatic effect on Southwest's OTP. In periods of marginal visibility, the model effectively manages arrival rates, and reduces the risk of diversions and delays. The model considers many different factors in arriving at the optimal solution including crew factors, maintenance routing, curfew, mission requirements, aircraft capabilities and capacity. Designed and built entirely in house at Southwest Airlines and was customized to fit the business needs and operating philosophy of Southwest Airlines.
The Social Connected Traveler
Daniel Stecher - Lufthansa Systems
Social media is there since more than ten years. Facebook was created 2004. 2016, most airlines still use social media primarily for marketing purposes, to play games with passengers or promote new routes. Contrary to airline passengers who massively use social media to complain about bad airline services, disruptions like late departures or arrivals, cancelled flights, missed bags, missed connections or bad inflight experiences. Airlines have to react then on these social media published complains and get back to the passengers. It's time to swap from reactive to proactive disruption management in general and consideration of the social connected passenger in particular. What's the impact of a happy airline passenger who positively tweets to his thousands of followers? Airline operations control can focus on potential disruptions with affected social connected travelers.
Towards Improved Decision Making through Data-Driven Operations and Maintenance Applications
Christian Strottmann Kern, Cyntia Cristina de Paula - EMBRAER
For the past 10 years, Embraer has been investing in Research & Technology (R&T) efforts to develop data analytics applications. These efforts have evolved from Prognostics and Health Monitoring (PHM), through Integrated Vehicle Health Management (IVHM) applications all the way to data-driven decision support systems focusing on the integrated operation of the aircraft within the aviation ecosystem. This presentation will outline that continuous evolution, showcasing a brand new service launched in 2015 related to flight operations and crew training and some of the latest solutions under R&T. A specific case around line maintenance planning will be presented in greater technical detail, providing an example of how an integrated view of an operator's resources can lead to improved decision making and more efficient operations.
Uncovering Delay Distribution Patterns Using Machine Learning Algorithms
Elham Boozarjomehri - Boeing Canada - AeroInfo
Disruptions greatly limit the ability of airlines to operate efficiently and effectively. Disruptive events come from a variety of sources and range greatly in magnitude making them challenging to predict and prepare for. In this study, we examined how big data and unsupervised machine learning algorithms could be used to shed light on the nature of airline disruptions. More specifically, we aimed to identify typical geographical patterns of daily delays experienced at airports in the continental US. Analysis of over 25 million flights revealed that while each day's delay distribution profile is unique, patterns could be generalized into a small number of clusters or recurrent geographic patterns. In an effort to identify the characteristics of each cluster, prediction models which correlated significant weather phenomena to the geographic patterns were developed. Early results were promising with models selecting the correct cluster with up to 80% accuracy.
Understanding Airline Performance through the Airline Benchmark Rating
Ryan Leick - Utah Valley University
The airline benchmark rating (ABR) is an index of an airline's contribution to impacting operational performance. Airline performance can be evaluated using a metric of "expected delay" to represent the ability of an airline to either out-perform or under-perform the historical delay given similar circumstances. Expected delay is derived from a regression analysis of multi-sourced flight records, ASDI, and ASPM data. Airlines are rewarded or penalized according to the "expected delay minutes" gained or lost from each flight. Using this methodology, on-time performance can be compared across the system and between airlines regardless of weather and airport congestion.
Wag the Tail
Tomas Larsson - Jeppesen
Crew planning is generally more complex and more constrained than aircraft allocation. Still, general practice in airline planning is that aircraft rotations are input to crew planning. Here we discuss how that process can be turned around and we show what can be achieved by an integrated crew and tail planning process where crew planning suggests re-timings and connections in aircraft rotations. Finally, results from applying those methods at various airlines are provided.
Weather Impact Analytics: Data-Driven Strategies for Managing Airline Operations during Adverse Weather
This paper presents some novel methods, tools, and opportunities to advance towards data-driven solutions for more informed and efficient airline operations during adverse weather. Through the use of objective methods for identifying and collecting similar events by weather, impacts, and air traffic actions, we will demonstrate how “outlier behavior” – both positive and negative – relative to similar weather-induced irregular operations scenarios can be identified, then internalized by the operation for decision-making with future impact events. We will discuss how weather-impact analytics can evolve towards predictive decision support of real-time airline operations and how this information could potentially integrate into an airline’s broader operating environment.