2016 INFORMS Annual Meeting Program

203 / 561

INFORMS Nashville – 2016
203
decision-making. A mixed Logit model with two latent classes, myopic and
looking-ahead route choice, is specified and estimated. Factors influencing
looking-ahead behaviors such as information bias, experience, network
complexity, and cognitive load are studied.
4 - The Initial Condition Problem With Complete History Dependency
In Learning Models For Travel Choice
Yue Tang, UMass Amherst,
Missing initial observations in longitudinal data can lead to inconsistent
parameter estimates in learning models for travel choice due to the complete
history dependency of choices of such models. This study proposes the maximum
simulated likelihood (MSL) and multiple imputation (MI) methods to address this
problem, and examines their efficacy and efficiency using an instance-based
learning (IBL) model. Monte Carlo simulations with synthetic data are carried out
to verify that the true parameter values are retrieved. An experimental dataset of
repeated binary route choice is used to illustrate the empirical applicability of the
methods.
MC61
Cumberland 3- Omni
Rail Safety and Risk
Sponsored: Railway Applications
Sponsored Session
Chair: Xiang Liu, Rutgers, The State University of New Jersey,
Piscataway, NJ, United States,
1 - Analysis Of Railroad Accident Investigation Reports Using
Probabilistic Topic Models And K-means Clustering
Trefor Williams, Rutgers University, New Brunswick, NJ,
United States, , John Betak
Railroad accident investigation reports from the National Transportation Safety
Board in the United States and the Transportation Safety Board of Canada were
analyzed using Latent Dirichlet Allocation (LDA) and k-means clustering. The
analysis suggests that recurring accident themes are track defects, wheel defects,
grade crossing accidents, signaling issues, train crew fatigue and switching
accidents. The Canadian analysis additionally highlighted accidents related to
bridges and track drainage. The LDA analysis of the accident reports will also be
contrasted with the results from an LDA analysis of the text fields of the FRA
Railroad Equipment Accident database.
2 - Statistical Modeling Of Freight Train Accident Estimation
Zhao Wang, University of Illinois at Urbana - Champaign,
Urbana, IL, United States,
Accurate train accident rate estimation is critical in railway safety and risk
management. In current methodology, the accident rate estimation methodology
is based on a single exposure variable and could be improved. Utilizing statistical
distribution modeling methods, a new train rate predictive model is build that
accounts for more complex factors.
3 - Simulation-based Risk Analysis Model For Optimizing Rail Flaw
Inspection Frequency
Xiang Liu, Rutgers University,
A simulation based risk analysis model is developed to quantify broken rail caused
derailment risk on U.S. freight railroads. The model can be used to determine
optimal inspection and maintenance scheduling.
MC62
Cumberland 4- Omni
Market Structure, Competition and Constraints
in the Airline Industry
Sponsored: Aviation Applications
Sponsored Session
Chair: Martin E Dresner, University of Maryland-College Park,
R H Smith School of Business, College Park, MD, 20742, United States,
1 - Excess Inventory As A Market Entry Deterrence Mechanism:
Evidence From The Us Airline Markets
Chen Zhou, University of South Carolina,
Inventory may have strategic value because it is necessary for production and
consumer satisfaction that are crucial for firms to achieve competitive advantage.
In service industries, inventory is often a result of extra capacity, which play a
critical role in customer service. In this research, we examine the strategic role of
inventory beyond cost efficiency and customer service, using data on the domestic
airline industry in the United States. We find that extra inventory discourage
market entry on major routes and the intensity of market competition
strengthens the relationship between inventory and market entry.
2 - Modeling Multimodal Network Equilibrium For Unregulated
Intercity Travel: An Equilibrium Problem With Equilibrium
Constraints Approach
Bo Zou, University of Illinois at Chicago, Chicago, IL, United
States, , Mohamadhossein Noruzoliaee, Lili Du
This paper proposes a new approach to modeling network competition for
multimodal intercity transportation. We view the interactions of transportation
operators and travelers as a Multi-Leader-Multi-Follower game, which is modeled
as an Equilibrium Problem with Equilibrium Constraints characterized by a series
of transportation operators’ profit maximization problems. To solve the EPEC, we
propose an algorithm that combines the Gauss-Seidel diagonalization method and
the relaxation method. The validity of the EPEC model and the algorithm is
demonstrated by applying them to a small network as well as a large network
representing the US Midwest.
3 - Costs And Benefits Of ‘Open Skies’ In The East African
Community (EAC)
Megersa Abate, Swedish National Road and Transport Research
Institute,
This paper investigates the economic effects of open skies in the East African
Community (EAC) member countries. We find two important results. Firstly, fare
levels are lower and departure frequencies are higher for countries with the most
liberal air services agreements (ASA). Secondly, our welfare estimate shows that
liberal ASAs have had significant economic impact in the order of $400 gain per
passenger. The paper also makes normative analysis of the role of liberal policies
on the long-term supply side responses in the EAC’s air transport market with
respect to airline output, market structure, fares and the position of national
airlines.
4 - The Impacts Of Merging And Acquisition On Route Entry
Behavior: An Empirical Test Of The Contingency Framework In
The US Domestic Airline Industry
Li Zou, Embry-Riddle Aeronautical University, ,
Janani Thiagarajan
This paper examines the impacts on route structure development of the merger
and acquisition (M&A) within the airline industry. Several contingent factors
considered include the timing of the M&A in the economic cycle, the financial
strength of the merging airlines, and their inclination toward merger activities,
and similarity in terms of cost structure, operation scale, market strategy, and
network structure. The estimation of the model is based on data drawn from the
US domestic airline industry over the 1993-2011 period. A total of 14 airline
mergers are covered in the analysis using data collected from Form 41, DB1B
Market Data, T100 Domestic Segment Data, and other supplemental sources.
MC63
Cumberland 5- Omni
Robust Planning in Air Transportation Systems
Sponsored: Aviation Applications
Sponsored Session
Chair: Adan Vela, Technical Staff, MIT Lincoln Laboratory,
244 Wood Street, Lexington, MA, 02420-9108, United States,
Co-Chair: James Jones, Technical Staff, MIT Lincoln Laboratory, 244
Wood St., Arlington, MA, 02420, United States,
1 - Inferring The State Of The Air Transportation System Using Mobile
Phone Data
Eric Feron, Georgia Institute of Technology, Atlanta, GA,
United States, , Aude Marzuoli,
Emmanuel Boidot, Helene Piquet
The Air Transportation system is becoming increasingly congested, leading to
perturbations that propagate through the entire airspace, disproportionately
impacting passengers. Most performance metrics in aviation are flight-centric,
simply because passenger data belongs to airlines. Yet, using anonymized mobile
phone location data, we can analyze passenger flow movements between airports,
under nominal and degraded conditions. Once passengers are identified amongst
mobile phone users, a matrix representing passenger flow movements between
airports is extracted. We study the impact of a bad weather event on passenger
flow movements, connections and wait times at airports.
MC63