|
AGIFORS Annual Symposium 2007 |
|||||||||||
|
Technical Program ( Based on accepted abstracts to date* )Monday - 01 October 2007 Expanding Beyond OR Dash
Continuous Improvement at United Airlines United Airlines has invested in assembling a comprehensive team focused on Continuous Improvement under one umbrella. The basic elements of the program include process improvement methodologies like Six Sigma and Lean, Industrial Engineering practice, Operations Research techniques, data mining and analytics concepts, all built on a robust program management infrastructure. We'll cover the infrastructure required in assembling such a practice, the purpose and objectives of this group and lessons learned on recent projects and some initial results. The approach taken in achieving goal alignment across business units, identifying and implementing improvements across the company to ensure sustainable results will also be discussed. Multi-Objective
Approaches for Robust Airline Scheduling The operational performance of an airline schedule is the result of a complex interaction between the schedule and its operating environment. A robust schedule is one that anticipates the stochastic nature of the operating environment and reduces the influence of disturbances on its operation. Evidence in the literature shows that the robustness of airline schedules is influenced by multiple features in the schedule, defined as robustness objectives. The construction of robust airline schedules should therefore be considered as a multi-objective optimisation problem that generates schedules with a good balance between the individual robustness objectives that maximise the operational performance of the schedule. We present time window approaches for incremental and integrated multi-objective improvement of robustness objectives in airline schedules. We applied single- and multi-meme memetic solution approaches enabling us to generate diverse sets of high quality trade-off schedules that balance the individual objectives. A large scale simulation was carried out to estimate the operational performance of the individual trade-off schedules. A sensitivity analysis of the simulation results was undertaken to estimate the individual and simultaneous influence of the robustness objectives and the impact upon the operational performance. Our work has enabled us to better understand the mutual interaction of robustness objectives and their individual and simultaneous influence on the schedules' operational performance. A Column Generation Algorithm for
Choice-Based Network Revenue Management In the last few years, there has been a trend to enrich traditional revenue management models built upon the independent demand paradigm by accounting for customer choice behavior. This extension involves both modeling and computational challenges. One way to describe choice behavior is to assume that each customer belongs to a segment, which is characterized by a consideration set, i.e., a subset of the products provided by the firm that a customer views as options. Customers choose a particular product according to a multinomial-logit criterion, a model widely used in the marketing literature. In this paper, we consider the choice-based, deterministic, linear programming model (CDLP) of Gallego et al., and the follow-up dynamic programming (DP) decomposition heuristic of van Ryzin and Liu, and focus on the more general version of these models, where customers belong to overlapping segments. To solve the CDLP for real-size networks, we need to develop a column generation algorithm. We prove that the associated column generation subproblem is indeed NP-Complete, and propose a simple, greedy heuristic to overcome the complexity of an exact algorithm. Our computational results show that the heuristic is quite effective, and that the overall approach has good practical potential and leads to high quality solutions. Tuesday - 02 October 2007 Using Simulation to
Appreciate Customer Experience at Dubai Airport As Emirates, and Dubai, continue their rapid expansion we have used simulation to understand the impacts to passengers and baggage at our hub. We know our operation intimately and some of the constraints around facilities. How do we effectively use that knowledge to create insights from a customer's perspective? While we wait for a dedicated Emirates terminal, simulation has answered many questions of how we can operate with the authorities and other airlines to influence informed decisions; to try and minimize the discomfort to our customers; and keep our hub operation on schedule. An Optimized Crew
Scheduling Proposal with Optimization Tracking and scheduling duties in Delta Airlines' Operation and Control Center are currently performed separately. Many of the tracking functions have been incorporated into the real time Crew Reroute Optimization System. The scheduling functions are performed manually with automated tool assists. Delta Airlines has been using the Crew Reroute Optimization System in production since June of 2002 for pilots and December of 2002 for Flight Attendants. Trackers define problems to a Server from a GUI and the Server then inputs the data for each problem to an Engine. The Engine solves the problem based on its parameterized legality and cost model and then sends the optimized solution including new (unpublished) pairings to the Server. Scheduling functions involve assigning available crew members to new (unpublished) pairings produced by the Trackers. An automated optimization crew scheduling approach is proposed. Methods for combining the tracking and scheduling functions are discussed. A prototype is described and some initial test results are presented. Optimized Control of the Number and
Scale of Group Passengers Group passenger is part and parcel for Asian aviation market. But the problem of group passenger control is rarely involved in revenue management. We design a optimization method to group passenger control. Each group passenger has three peculiarities: 1 each group must be all in the same flight, 2 each group books to airlines, 3 the price to each group is according to booking time and group scale. We set up a mathematic model which includes single passenger demand and group passenger¡¯s price and scale. The model is a 0-1 programming. We design a bound-branch arithmetic to solve the 0-1 programming. ROAR: Robust Operational Aircraft
Routing We developed an approach to ensure that a schedule is maintenance and operationally friendly. The results of fleet assignment models (FAM) often do not meet airlines’ 3-4 day maintenance requirements. We take FAM output of a fleet assignment model and generate intra-fleet aircraft swaps to meet 3-4 day maintenance requirements. Furthermore we look for opportunities to increase schedule robustness by considering operational characteristics such as crew connections and eliminating tight turns. We observe that we can further improve maintenance and robustness results by swapping across fleet types. In order to evaluate the revenue impacts of inter-fleet swaps, we develop a fast O&D revenue management model. We combine this with a model to identify the best combination of potential swaps. Starting with a typical FAM output, the swapper can improve maintenance feasibility, robustness and profitability. Strategies for Accurate Passenger
Volume Forecasting Passenger volume is one of the principal drivers for a commercial airline's operation. Increased accuracy in daily passenger volume forecast leads to efficient management of service staff, leading to better customer experience as well as higher cost saving. We apply time-series method with an innovative twist in selecting optimal weights to forecast daily inbound and outbound passenger volume from large stations. Daily local, terminating and connecting passenger volume are forecasted with high degree of accuracy. Taking the forecast one step further local, terminating or connecting passengers traveling on each flight can also be forecasted by proper utilization of historical information. Determining the
Passenger Value of Departure and Arrival Times Using discrete choice methods of measuring passenger utility, the issue of determining the value passengers assign to alternative departure and arrival times – the Time of Day Demand problem – is examined. A logit formulation is described, as is common with discrete choice models, but which departs from the standard construction in two significant ways. First, a discontinuous passenger utility function is asserted, with an indifference window within which the passenger has no time-based utility difference, one utility description for departures prior to the indifference window, and a different disutility for departures beyond the indifference window. This yields a logit model wherein the utility description is no longer linear in the parameters. Second, a separate probability distribution, describing the incidence of ideal departure times in the population, is used as a mixing function to create a mixed logit distribution, from which market share and price premiums can be calculated. Using data from a stated preference survey conducted by Boeing, both the utility function and the mixing distribution are estimated, using partial maximum likelihood techniques for the utility and a sine-cosine decomposition for the mixing distribution. Sample calculations of share and price premium will be presented. Research on the Optimization Control
Method of the Quantity of Aperiodic Ticket Aperiodic ticket, an effective method of adjusting demand and resolving the inconsistency between supply and demand, has been put into practice in the domestic and foreign airlines. An optimization control method of the quantity of the aperiodic ticket has been addressed in this paper. Using the expected revenue and EMSR, we promote two propositions and set up a model to optimize the quantity of the aperiodic ticket. Then we present a heuristic arithmetic and program by MATLAB language. Finally, we analyze the relations among the quantity and the price of aperiodic ticket and the total revenue by an instance. Compared with the revenue calculated by EMSR without aperiodic tickets, the total revenue raises 5.05% when the quantity of the aperiodic ticket is optimized, and 5.28% when the quantity and the price are optimized at the same time. This shows that the optimization of the quantity of the aperiodic ticket can increase flight revenue of airlines. Role of Enterprise
Analytics at United Airlines This talk will cover recent contributions of R&D at United Airlines in the following areas: Block Planning – Modifications to Block Planning approach to enhance Customer Experience. Schedule Planning – Ground Time Optimization as well as close in Re-fleeting Revenue Management – Buy down aware revenue management algorithms Airport Operations – Airport Manpower Optimization. Thursday - 04 October 2007 The Internet Price
Effect Continuing the themes discussed in two previous AGIFORS Presentations, the author will present the results of his doctoral research quantifying the impact of Online search on the fares paid by customers across various markets over a five year period. The regression analysis focuses on prices paid by distribution channel, controlling for several trip, market and individual customer characteristics, and for the Opportunity Value of the seat the customer actually used, as measured by the EMSR of the flight.
FFP: Frequent Forwarder Program – Exploring Cargo Loyalty
Management Widely accepted to be an industry that relies heavily on Business to Business (B2B) relationships, the air cargo business has witnessed relatively scant attempts to innovate in the field of cargo loyalty management. Often argued to be impractical, not feasible, or not desirable, the concept of cargo loyalty management should be explored owing to the proven results of corporate loyalty programs in other fields. This address delves into the multiple opportunities to institute “frequent forwarder programs” drawing from supporting structural trends in the air cargo transportation and logistics supply chain by examining and confirming the prospect and benefits of total “Carrier to Forwarder to Shipper” loyalty management (or B2B2C). Economic Analysis of Spider Web
Airline Networks The distinct network organization, management, service and operating characteristics of US Southwest Airlines comparing to other airlines are key elements of its success. Spider web networks as the network organization has been paid more attention to. The paper analyzes the structure and relation with the spider web, builds the assignment model of spider web airline networks and examines the economy of spider web airline networks. Designing a Cargo Hub
using Simulation In the near future KLM Cargo will move to a new location. This involves a big investment, but at the same time it is a good opportunity to redesign the freight building according to the latest insights and circumstances. We have developed a simulation model of the processes in the building which supports the decision making by giving insight in the effects of certain design decisions and the infrastructure needed. The simulation model is built using a modular approach. Its input consists of flight and trucking schedules in which growth scenarios are included. State of the Airline Industry There are significant variations in the state of the airline industry around the world. What are these variations, how have they come about and what is likely to happen in the future? What have been the direct influences of fuel price rises? What are the ongoing effects of low-cost airline proliferation? What will world fleets look like in 10 years time? Schedule Design and
Fleet Assignment with Demand Supply Interactions The paper solves schedule design and fleet assignment problem while taking care of various demand-supply interactions. Scope of the paper includes both unconstrained and constrained demand. Paper provides a method to incorporate the change in unconstrained itinerary demand as a function of supply and the approach is less subjective and hence more scalable as compared to existing methods. Paper also discusses a new approach to handle spill and recapture. The resultant model is a mixed integer non-linear model that has been converted into a mixed integer linear problem. The paper illustrates the utility of approach with the help of examples. Solving an Irregular Airline
Operation using Multi Objective Micro Genetic Algorithm Several approaches such as linear programming, network modeling, greedy heuristic and decision support system are well-known approaches in solving irregular airline operation problems. This paper presents an alternative approach based on Multi Objective Micro Genetic Algorithm. We introduce the concept of Multi Objective Micro Genetic Algorithm as a tool to solve irregular airline operation, combine and reroute problem. An experiment was carried out to observe the convergence behavior and running time until the convergence is reached. The test data is a simulated flight schedule, based on Thai airways. The preliminary result indicated that, after running the model 10 times and each time contained 100 epochs or 100 generation, we successfully discover a set of optimal solution within 100 generations of each running time that takes approximately 8 second. Reinforcement Learning in Irregular
Flight Operation While most works on the irregular flight operation problem seek to obtain solutions by mathematical models or network theory, few attempt to apply reinforcement learning concept. This may stem from the fact that most machine learning algorithms are not feasible when real-world applications demand the solution in almost real-time. Q-learning, a more recent form of reinforcement learning, is an algorithm that offers solutions in real-time because it does not need the environment model. In this paper, we apply the Q-learning algorithm to solve irregular flight operations at Thai Airways International. Preliminary test shows promising results compared to the solutions from the existing system. Friday - 05 October 2007 An Overview of
Integrating Origin-Destination Revenue Management This paper presents an overview of the fleet assignment, schedule development and revenue management processes airlines typically use to produce, fleet and manage the seat inventory of their schedule. In addition, this paper presents a methodology for incorporating O&D network effects into the fleet assignment process. The methodology uses a decomposition strategy to combine a modified version of a leg-based fleet assignment model (Leg-FAM) with the network flow aspects of probabilistic O&D revenue management. By decomposing the problem, the nonlinear aspects of the O&D market effects and passenger flow are isolated to the O&D revenue management process and incorporated into the fleet assignment process using linear approximations to the network revenue function. To illustrate the O&D fleeting concept and its benefits, we apply this concept to a Demand Driven Dispatch (D3) scenario where near-term fleeting changes improve the match between O&D passenger demand and available aircraft capacity close to the day of departure for American Eagle Airlines. Airline Planning and Operations
Tutorial
Call for Technical PapersThe 47th Annual Symposium of the Airline Group of the International Federation of Operational Research Societies (AGIFORS) features a comprehensive program covering the whole spectrum of airline operations research. This year’s symposium will be held from 30 September – 05 October 2007 in Bangkok, Thailand. Contributions to the technical program are solicited in the following areas, but are not limited to, flight scheduling, pricing, yield management, maintenance and engineering, crew, ground resources, and finance. Papers covered by active Study Groups, as well as other topics not covered by active Study Groups, are sought. Airline, consultant, academic, and industrial research representatives are all encouraged to attend. Presentations are 30-45 minutes, inclusive of minutes for Q&A. Symposium proceedings will be published by January 2008. *Your registration discount is based
on the conditional acceptance of your proposed presentation for the
symposium technical program, and only applicable to the primary
presenter. Deadlines and Requirements Abstract Submissions - August 31, 2007 (extended) 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 - September 15, 2007 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 2007:
Inquiries about Technical Program Tim
Jacobs |
||||||||||
Welcome | Agenda | Registration | General Information | Hotel Information | Host City |
|||||||||||