AGIFORS Annual Symposium 2009
Delivering Operations Research Based Innovation

Welcome

Agenda

Registration

Technical Program

General Information

Hotel Information

Social Activities

Official Sponsors

Host City

 

Technical Program

These submitted abstracts have been conditionally accepted for inclusion in the AGIFORS Symposium agenda. Final acceptance is contingent on 1) receipt of complete presentation on/before 20 September 2009, and 2) approval of presentation content by the AGIFORS technical committee.

Airline customers online search and purchase for discounted fare products
Misuk Lee and Laurie Garrow, Georgia Tech, David Post (Sigma Zen)


This study analyzes how individuals search for information about air ticket attributes and prices using data from a website that sells tickets at low prices to travelers with a high degree of travel flexibility. Markov-based models are used to capture the dynamics of customers’ searching and purchasing behaviors. Empirical results reveal that what visitors saw in previous site visits clearly influences future revisit and purchase probabilities, as well as search intensity within sessions.

 

A Paradigm for Scheduled Flight Services
Jon S Kyle, S8 Cargo Inc.


A novel paradigm provides a way to categorize origin-to-destination flight services. The paradigm derives from the “efficient frontier” defined by Markowitz in his groundbreaking 1952 paper that launched financial portfolio theory for which Markowitz won the 1990 Nobel Prize in Economics. The paradigm gives rise to an algorithm for generating origin-to-destination flight services with astonishing efficiency. The algorithm permits direct generation of a set of “best services” without ever considering services that are not members of the set of best services. The set of “best services” is equivalent to Markowitz’ “efficient frontier”. Products using the algorithm will be briefly demonstrated.
 

A new methodology for solving the no purchase problem in choice based RM
Laurie Garrow, Mark E Ferguson and Melike Meterelliyoz, Georgia Tech

 

Through cleverly redefining customer arrival rates and using historical market share, we leverage the independence of irrelevant alternatives property of the MNL model to calculate a non-zero no purchase utility. Our algorithm solves for MNL parameters using one iteration of the EM algorithm, effectively reducing computational times from hours to seconds. Most important, MNL parameter estimates are both intuitive and stable across datasets.
 

Baggage Capacity and Demand Management Becomes even more Sophisticated
Desmond Di Wang, Northwestern University.  Additional Authors: Diego Klabjan (Northwestern), Robin Riedel (AC), Steven Truttenbach (AC), Jean-Francois Page (AC)


Load of an aircraft consists of passengers, cargo, and baggage. While sophisticated forecasting systems exist for the first two contributors, this is not the case for baggage despite significant costs associate with baggage mishandling. It is not uncommon that on the day of the operations the aircraft is overloaded with excessive baggage, which leads to dire consequences for both passengers and the airline. It is thus desirable to detect such situations early during the booking process. We present methodologies behind a system for baggage load forecasting any number of days before the day of operations. We will discuss the data gathering process and the underlying implementation of the system.

Bounds for network revenue management
Kalyan Talluri, Universitat Pompeu Fabra in Barcelona


Tight upper bounds for network revenue management are of interest in themselves as they show how close a method is to optimum, and also as the bounds often lead to new control methods. In this talk we compare some known bounds and present new bounds for network revenue management, that are provably tighter than some of the existing bounds.
 

Estimating OD Demand with New Solutions to an Old Problem
Roger A Parker, Virtual M1nds


The estimation of ‘true’ OD demand has long been a central issue for airline analysts. This presentation will survey the current state of the art, and then offer two recent advances made possible by the application of discrete choice theory and modern simulation modeling. First, a discrete choice ‘gravity’ model combined with a nested logit structure will be described which applies the classic trip-generation/trip-distribution construct to build a market-level OD demand model for the US. Second, the concept of NCAD – Network and Channel Adjusted Demand –will be introduced. Using agent simulation, it holds promise a significant advancement for the field.

 

Equitable and Efficient Coordination of Traffic Flow Management Programs
Douglas Fearing, Massachusetts Institute of Technology
Anna Valicek Medal 2009 Finalist
 

We propose two optimization formulations balancing equity and delay for the multi-resource TFM scheduling problem. To evaluate these models and compare them to the current approach, we develop a metric for schedule fairness derived from highly-successful properties of RBS. Through regional and national scenarios derived from historical data, we demonstrate that both models lead to improved efficiency while maintaining an equivalent level of fairness as current practice.

 

Improving schedule profitability using flight retiming
Sergey Shebalov, Sabre Inc.


We present an optimization models that improves schedule profitability by retiming flights at an airline hub. The model employs logit O&D demand forecasting methodology and therefore accounts for demand share changes caused by retiming for both local and connecting markets. Additional revenue is due to improved quality of existing itineraries as well as new connection opportunities. We propose to employ this model at the network planning stage as we make sure that operational constraints such as aircraft count and flow balance are satisfied, and so the modified schedule is feasible for fleet assignment. Because problem size can become formidable, we develop an iterative approach that gradually improves schedule structure and keeps memory usage and run times at reasonable level. Our experiments on a real airline data show up to 3% improvement in revenue opportunity.
 

Integrated Digital Marketing And Revenue Management
Kevin Geraghty, Revenue Research Inc.


This paper explores the synergies between digital marketing and revenue management by way of case studies including a dynamic pricing implementation that drove $50 million in revenue enhancement in 3 months. It also provides a sneak peak into future tools that Google is providing that will significantly change price delivery on the web.

Profit Maximization of Multi Stops Operating Model
Mohammed Salem, Felix Airways


The competitive environment of aviation industry is driving most airlines to re-evaluate their strategies to survive, and many strategies are setup either to reduce the cost or to improve the profits as e-ticketing, simplifying the Business, e-freight to meet the challenge of Low Cost Carriers in their markets. This paper is addressing a new approach for maximizing profits in Multi-Stop Operating Model in long haul operation of legacy airlines based on demand passengers of multi destinations, corresponding market fares and cost of available ton kilometers of the assigned aircraft. Linear program is developed and many constrains are set up. The profit is maximized by solving operating route/ arranging the right number of frequencies, load factor per sector and cabin load factor for each respective destination. Based on these results many scenarios are implemented to define the right aircraft to operate and the most profitable routes to select in the network. The program can be further extended to study the effects of new destinations on the existing routes / network. Meanwhile, market opportunity is another important factor to develop a decision matrix with a profit to define the complete picture of the airline network.

 

Revenue Management Models and their Estimation
Kalyan Talluri, Universitat Pompeu Fabra in Barcelona

 

We report results of a large scale empirical study of bookings data across four RM industries (network airline, low-cost, cargo and hotel) examining if the assumptions of RM models are justified. The main technical difficulty is that we cannot observe no-purchases. We present alternate models that better fit data and are also tractable. We present alternate estimation procedures and report simulation experiments on their performance.

 

Synchronized Aircraft and Crew Rotations
Tore Jenner , SITA

 

To achieve a robust airline operation where delays are not propagated crews should ideally follow the aircraft flow. For reasons like industrial contract limitations, organization, planning methods and computer systems this cannot always be obtained and is particularly difficult for network operators. The presentation elaborates on the difficulties and shows how optimization can be used to synchronize aircraft and crew rotations to improve operational robustness.

 


Scheduled Maintenance Planning for Per Seat On Demand Air Transportation

Gizem Keysan, Georgia Institute of Technology
Anna Valicek Medal 2009 Finalist
 

We address planning the scheduled maintenance of per-seat, on-demand air transportation. At the tactical level, we determine the maintenance capacity, i.e., the maximum number of jets that can be maintained on a given day. We model these decisions with an integer program (IP) and use an optimization-based local search algorithm to solve the IP. The specific jets to be maintained on a daily basis are determined at the operational level with a decision support tool based on a two-phased IP approach. Finally, we develop an integrated framework to capture the interaction between the operational level maintenance decisions and flight scheduling.

 

The Ergonomics of Corporate Decision Making
Stefan E Karisch, Jeppesen


At Jeppesen, operations researchers are routinely called upon to provide analytic support for high-value decisions under conditions of great uncertainty. Such decisions concern for example the corporate investment portfolio, product launches and inventory control. The methodologies deployed comprise multi-objective optimization, integer and nonlinear programming, statistical analysis, and simulation. While analytic techniques provide valuable supporting information for decision-makers, the decision-making process itself turns out to determine the success or failure of our efforts. In this talk, I will describe critical success factors of the decision-making processes at Jeppesen. They include, besides a flexible and adaptable technical methodology, a human-centered approach.
 

Understanding the Relationships between  Price, RM Controls and Scheduled Capacity
Tim Jacobs, American Airlines.  Additional Authors: Barry Smith (Barry C. Smith Inc.) and Richard Ratliff (Sabre Holdings)

 

The relationship among pricing, RM controls and capacity directly impacts the profitability potential of an airline’s schedule.  Too much capacity in a market leads to wide open RM controls and bargain basement prices to stimulate demand to fill the available seats at unprofitable yields.  On the other hand, a schedule that does not supply enough capacity can lead to higher prices and stringent controls that invite competition.  Understanding the relationship between the scheduled capacity, pricing structure and RM controls gives an airline the opportunity to better align scheduled capacity, adjust pricing structures and manage the seat inventory to maximize profitability.  This paper presents one methodology for evaluating the impact of scheduled capacity and pricing structures on overall network revenue.  This paper presents the macro level metric for evaluating the quality of a given pricing strategy and guiding a search algorithm to identify the optimal an optimal alignment between pricing structure, scheduled capacity and the RM process.

 

Deadlines and Requirements  

Abstract Submissions - 15 September 2009

Please send all submissions via the online form.

Submit abstracts online (maximum 100 words). Include a title and identify the corresponding author.  For each co-author, include full name, affiliation, complete address, telephone number, FAX number, and electronic mail address or URL.

Paper/Presentation Submissions – 20 September 2009

Corresponding author provides both an electronic file submission of the presentation/paper for inclusion on the CD-version of the Proceedings and a photo-ready copy for the printed Proceedings.

Please note the following conditions for inclusion of your paper in AGIFORS 2009:

  • At least one author of an accepted paper must attend the Symposium to make the oral presentation.
  • Each attendee must register and pay the registration fee.
  • Each attendee is responsible for making all hotel arrangements.

Inquiries about Technical Program

Tim Jacobs
AGIFORS Past President, and Technical Director
Phone: +1 817.967.1841
E-mail: tim.jacobs@aa.com

 
Welcome | Agenda | Registration | General Information | Hotel Information | Host City