Log in

SSP 2024 Technical Presentation Abstracts

Tentative content, subject to review of draft presentations.

Additional abstracts to be added as received.

Swap-based Heuristic for Aircraft Routing

Nitin Srinath, United Airlines

In airline network planning, aircraft routing is an important problem. This task is essential not only during the early stages of schedule planning but also as the day of departure approaches. Ensuring that each aircraft maintains a feasible flying schedule is crucial, particularly to facilitate regular maintenance opportunities. Our project concentrates on refining the initial schedules to better accommodate maintenance needs. To begin with, we have introduced a novel scoring system designed to gauge the compatibility of a schedule with established maintenance regulations. Following this, we have developed a simple swap-based heuristic that leverages this score to enhance the lines of flying for individual aircraft. The implementation of this routing algorithm has significantly improved maintenance scheduling, with certain aircraft fleets witnessing up to a 70% enhancement in maintenance performance.

From Vision to Reality: Our Journey of Creating a Schedule Optimizer

Wendy Brave, KLM

Imagine a world where all the complexities of an airline schedule can be solved with a single touch of a button. It is not an easy job to include all the schedule dependencies into one optimizer which is also adopted by the business. Think Big, start small! Join us on an exciting journey where we tell you about our schedule optimizer, SOLAR. Discover the use case we are solving with our first version, learn more about the technology we use, and witness how we ensure that schedule planners embrace this optimizer. But that is just the beginning. We will also offer you a glimpse into the future, sharing our plans to expand SOLAR.

Integrating Gate Assignment Feasibility

Nicholas Etz, United Airlines

Airport gate assignment is an important factor in airline operations. This presentation introduces a model for determining whether a schedule can be gated at key stations. It assesses core gating considerations, including towing, as well as bespoke constraints like remote parking restrictions. When a full gating solution is infeasible, results are then used as a basis for retiming recommendations. By providing a quick spot check for gate assignment feasibility, the model helps integrate gating information throughout the network planning pipeline.

Simultaneous Optimization of Airline Network and Fleet with Evolutionary Algorithms and Data Pipelines

Mikhail Andriyanov, Andriyanov & Partners Mathematicians and Economists PartG

Choosing the right fleet is challenged by the curse of dimensionality, where each composition of aircraft types ideally requires its own optimized network. We solve this network and fleet planning task as a simultaneous optimization problem. Traffic forecast, demand stimulation, market share simulation, network traffic distribution and evolutionary optimization are integrated into an end-to-end data pipeline. Each run of the algorithm automatically outputs the optimal number of aircraft units, destinations, and frequencies for each fleet mix. Optimization can maximize any objective, including pure profit and profitable growth. Implementation as a Python pipeline allows accounting for any practical constraints and considerations, such as strategically operated routes, multi-stop flights and cabin configurations. We showcase the approach using open data from the U.S. Department of Transportation, where the network and fleet of a hypothetical airline based at San Francisco Airport are optimized under various objectives and conditions.

The Impact of Passenger Preferences for Schedule in an Airline Revenue Management Simulator

Emmanuel Carrier (speaker), Delta Air Lines & Peter Belobaba

As we develop a replacement for the Passenger Origin-Destination Simulator (PODS), we study the impact of passenger preferences for schedule on the performance of advanced revenue management (RM) methods. At the core of PODS sits the Boeing Decision Window model that is used to represent passenger preferences for schedule including preferred departure and arrival times and the attractiveness of shorter itineraries. We show that disabling these preferences shifts revenues from airlines with a superior schedule (more nonstops, shorter connections) to carriers that offer less convenient itineraries. Without modeling these preferences, simulation results substantially under-estimate the benefits of advanced RM algorithms such as network O-D control. A proper representation of passenger preferences for schedule will then be instrumental to provide a realistic environment to evaluate the performance of RM algorithms in the new simulator.

Predicting a Schedule's OTP and How that Impacts Schedule Build

Pascale Batchoun, Air Canada

The first step toward establishing a resilient flight schedule is to accurately predict its On-Time-Performance (OTP) and to evaluate the impact that both scheduling and operational changes could have on OTP. Air Canada has developed and deployed to production a machine learning-driven methodology that predicts system-level OTP key performance indicators (KPIs) while providing low level estimates for the block and turn durations at the flight level. Augmented with a simulation engine, the “OTP Schedule Optimizer” evaluates various scenarios to simulate disruptive and cascading delays, and their impact on our aircraft performance, passenger connections and crew flow. The system uses an optimization engine that leverages various outputs from the machine learning and simulation models, to recommend schedule changes to improve its OTP while minimizing the impact on passengers’ misconnections. We will shed the light on how we designed the system and share insights on performance KPIs, including simulation of operational changes, and what are some of the pain points highlighted in the schedule, and what actions to take in the Planning, Scheduling and operational windows.

The Acquisition of Regulatory Filings to Inform Pricing Decisions: Evidence from Airline Yield Management

Hengda Jin, Texas A&M University

This study examines airline yield management analysts’ acquisition of non-airline companies’ regulatory filings and the association of that acquisition with airlines’ pricing decisions. I find that yield management analysts’ EDGAR searches are positively associated with yields (i.e., the product of average airfare and passenger load factors). This finding suggests that information acquisition helps analysts forecast business travel demand and thereby maximize revenue through effective intervention in yield management systems. Decomposing information acquisition into different types of filings, I find that the positive association is driven primarily by the acquisition of accounting reports and varies with the characteristics and content of accounting reports. Specifically, the association is weaker when the accounting reports acquired are less readable, and it is stronger when the accounting reports acquired provide more quantitative forward-looking statements about business travel demand. Overall, the evidence enhances our understanding of how the customer base’s regulatory filings facilitate suppliers’ pricing decisions.

Optimizing Flight Itineraries in Hub-And-Spoke Networks

Luis Correa, Copa Airlines

In the complex world of airline scheduling, achieving optimal flight schedules is crucial yet challenging and often time-consuming. This presentation introduces an optimization model developed to tackle this challenge for a hub-and-spoke airline. The model attempts to optimize flight itineraries, considering key operational constraints such as block times, minimum ground times, maintenance requirements, slots, airport curfews, minimum connecting time at the hub, and flight separation criteria to avoid hub congestion. It focuses on maximizing network revenue by enhancing flight connectivity within banks, considering that not all arrivals can connect with all departures in each bank. Attendees will learn about the model's formulation, constraint integration, and the practical implications of its implementation in network planning and scheduling.

Supply-Demand Dynamics for Long-Term Aircraft Demand Forecasting

Younes El Jarari Charqui, Airbus

Traditionally, long-term aircraft demand forecasts predominantly hinge on macroeconomic indicators, often overlooking the potential that supply has in stimulating or constraining demand. The study challenges the conventional demand-driven paradigm by quantifying the extent to which, in the past, supply has played a key role in the realised passenger air traffic. A framework for introducing supply-side features, including aircraft production rates and airline capacity/network plans, is proposed to improve long-term forecasting models.

Schedule Reliability Optimization

Shahram Shahinpour (speaker) and Sureshan Karichery, Sabre Corporation

Reliability of airline schedule ensures smooth execution during operations, improves passenger experience and has direct effect on cost reduction and boosting financial bottomline. Therefore, it is imperative to measure on-time performance (OTP) metrics of the schedule and optimize them during network planning and schedule design tasks. In this talk, we first present a model to evaluate future schedule OTP at the network and flight level. Next, using evaluated OTP metrics and airline-specific business rules an optimization model is formulated and solved to produce the best solution that optimizes OTP and profitability objectives. The proposed optimization model is comprehensive with diverse set of rules and configurable target OTP levels. Benchmarks prove capability of proposed model in achieving planned OTP and financial targets.

The impact of Airport Congestion on Aircraft Sizing

Rohan Nanda, Airbus

The impact of airport congestion on aircraft upgauging is frequently contested, as there is considerable debate as to what extent this trend is driven by economic incentives versus airport slot constraints. As more airports are forecast to reach capacity limits in the future, there is a need to better understand this phenomenon. By analyzing flight data, airport capacity, and airline scheduling practices, we aim to dissociate between the two drivers in order to understand their relative contributions. In doing so, we gain insights into how this effect could be incorporated into forecast models, in order to better anticipate aircraft size requirements.

Powered by Wild Apricot Membership Software