2016 INFORMS Annual Meeting Program

114 / 561

INFORMS Nashville – 2016
114
SD59
Cumberland 1- Omni
Advances in Transportation Modeling
Sponsored: Transportation Science & Logistics
Sponsored Session
Chair: Tarun Rambha, University of Texas, Austin, TX, United States,
1 - Throughput Analysis For Horizontal Traffic Queues under
Safety Constraints
Mohammad Motie, University of Southern California, 1119 W
29th Street, Apt 7, Los Angeles, CA, 90007, United States,
We consider horizontal traffic queues (HTQ), where vehicle arrival and departure
locations are sampled from spatial distributions. We consider first and second
order car following models that guarantee no collision. HTQ is a state-dependent
queuing system, where the service rate depends on the configuration of vehicles.
We combine queuing- theoretic and dynamical systems tools, to provide novel
insights into the service rate dynamics and busy period distribution for HTQ.
These tools are used to compute bounds on throughput, which closely match
simulations. Our throughput analysis illustrates the interplay between car-
following behavior, road geometry, and arrival and departure statistics.
2 - Lost Demand And Redistribution In Bike Sharing Systems
Konstantina Mellou, Massachusetts Institute of Technology,
Cambridge, MA, United States, , Patrick Jaillet
Spatial imbalances in bike sharing systems often lead to unavailability of resources
(bikes or docks) and, as a result, lost customer demand. Our goal is to model
redistribution operations that will allow the company to improve its level of
service. An optimization approach is used, combined with decomposition and
heuristics, and the performance of our methods is evaluated with tests on real
datasets. Lost customer demand, which is often not considered since it cannot be
registered in the system data, is also taken into account.
3 - A Destination-based Algorithm For User Equilibrium With
Recourse Using Split Proportions
Tarun Rambha, University of Texas, Austin, TX, 6,
United States, , Stephen Boyles,
Avinash Unnikrishnan
When travelers receive en route information and select routing policies that
minimize expected cost, user equilibrium with recourse models can help predict
the resulting network state. We propose a new destination based algorithm to
solve such models using link proportions and compare its performance with
existing methods.
4 - Optimal Patrol Planning For Urban Parking Enforcement
Considering Driver’s Parking Behavior
Chao Lei, University of Illinois at Urbana–Champaign,
2063 S. Orchard St. Apt A, Urbana, IL, 61801, United States,
In the aim of designing an effective parking patrolling scheme for the parking
enforcement agency, we propose a bi-level optimization approach to help the
agency determine the patrolling schedule and routing plan in the upper level
while considering the drivers’ parking payment decisions in the lower level. Both
a mixed-integer formulation and a continuum approximation (CA) model are
developed. The numerical study shows that, due to its advantage in
computational performance, the CA approach provides a good alternative to
handle the large scale problems.
SD60
Cumberland 2- Omni
Emerging Data Analytics in Transportation Modeling
Sponsored: TSL, Urban Transportation
Sponsored Session
Chair: Xian-Biao Hu, Metropia, 1790 E. River Rd., Suite 140, Tucson,
AZ, 85718, United States,
1 - Contextual Driving Risks Analysis Using Individualized
Dynamic Smartphone
Xian-Biao Hu, Metropia Inc.,
Traditional driving risk study is usually based on crash history data and can only
be performed at aggregate level. Latest information and communications
technology allows individualized data collection, but most researches rely solely
on user GPS trajectories data but fail to consider other critical risk factors in the
surrounding environment that also contribute to crash risk. To bridge this gap, a
new approach that collects individualized driving behavior data from smartphone
GPS module, combined with geographical network information and dynamic
traffic conditions is presented to identify driving risk factors and evaluate driving
behaviors under various contexts.
2 - Put Bluetooth Data In A Good Use: A Case Study In
Tucson, Arizona
Shu Yang, University of Arizona,
Yao-Jan Wu
A Bluetooth based traffic data collection and analysis system is developed and
integrated into the regional transportation district network system. Fully utilizing
the Bluetooth-based data requires comprehensive data quality assurance
including data decomposition, imputation, and outlier detection. A case study is
conducted in Tucson to demonstrate a one-stop solution for Bluetooth-based
traffic performance measurement.
3 - Developing A Simulation Model Of A Tram Network By Using
Historical RFID Data
Yong-Hong Kuo, The Chinese University of Hong Kong, Hong
Kong, Hong Kong, Janny M. Y. Leung,
David S.W. Lai, Henry K.F. Cheung, Joshua Hiew
In this talk, we will present a real-world application that utilizes historical RFID
data for the development of a simulation model of a tram network. The historical
data about the tram locations are used to model the travel times of trams at
different times of the day. Our simulation model allows the tram company to
examine the impacts of different tram schedules on the service requirements and
other performance measures.
4 - Accident Impacts on Traffic Mobility In Concern Of
Network Features
Chenshuo Sun, Tsinghua University, Haidian District, Beijing,
100084, China,
It is hypothesized that there may be associations between intersection’s accident-
impact proneness and its location, as well as accident impacts and its origination.
To substantiate such hypothesis, four topological measurements are assigned to
intersections. Then, accident impacts specified in four aspects are quantified from
both macroscopic and microscopic perspective. This study employs one macro-
level model and three micro-level models. The results prove that intersection’s
accident-impact proneness is closely related to the network features of its
location, also accident’s infectiousness, network damage and delay are closely
related to the network features of its origination.
SD61
Cumberland 3- Omni
RAS Problem Solving Competition
Sponsored: Railway Applications
Sponsored Session
Chair: Lingyun Meng, Beijing Jiaotong University,
Problem Solving Competition
This session is reserved for the finalists of the RAS Problem Solving Competition
(PSC). The presenters and their abstracts won’t be determined until we finish the
judging process, which happens around mid-October. The selection committee
will identify the top three teams who will present their results during the session.
This year’s competition addresses how to route trains through a complex railway
network, with limited infrastructure capacity, while planning maintenance tasks.
SD62
Cumberland 4- Omni
Aviation Applications Section: Awards Finalists
Sponsored: Aviation Applications
Sponsored Session
Chair: Senay Solak, University of Massachusetts Amherst, 121
Presidents Drive, Isenberg 318, Amherst, MA, 01003, United States,
1 - Metaheuristics For Efficient Aircraft Scheduling And Re-routing at
Busy Terminal Control Areas
Marcella Sama, Roma Tre University, Rome, Italy,
This work improves a state-of-the-art optimization solver for the real-time
management of landing and take-off operations in a busy terminal maneuvering
area. The solver computes a good initial solution for the aircraft scheduling
problem with fixed routes, and then improves it by routing flexibility.
Metaheuristics based on variable neighborhood search, tabu search and hybrid
schemes are proposed. Experiments are performed on an Italian terminal
maneuvering area to simulate various types of disturbances. Solutions of
remarkable quality are computed within a short time.
SD59