2016 INFORMS Annual Meeting Program

284 / 561

INFORMS Nashville – 2016
284
2 - Fast Computation Techniques For The Stochastic On-time
Arrival Problem
Samitha Samaranayake, Cornell University,
We present a new technique for solving the path-based stochastic on-time arrival
(SOTA) problem. Our approach uses the solution to the policy-based SOTA
problem - which is of pseudo-polynomial-time complexity in the time budget of
the journey - as an efficient search heuristic for the optimal path. We also
demonstrate how path preprocessing techniques can be used for further
speedups. To the best of our knowledge, these techniques provide the most
efficient computation strategy for the path-based SOTA problem for general
probability distributions, both with and without preprocessing.
3 - Urban Freight Microsimulation: Evaluating Freight Parking
Behavior In New York City
Trilce Marie Encarnacion, Rensselaer Polytechnic Institute,
Jose Holguin-Veras, Johanna Amaya
As urban centers continue having increased demand for consumer goods and
services, the amount of freight trafficc and the associated negative externalities
increase. Previous studies have shown that carriers have to either cruise until
they find a parking space or illegally park in order to make their deliveries. Using
data collected from in-depth interviews of carriers as well as current traffic
conditions, a discrete simulation framework was developed to replicate the
parking behavior of trucks making deliveries to a case study area in Midtown
Manhattan. The goal is to provide insight into optimal freight parking policies to
improve freight systems performance in dense urban areas.
4 - Dynamic Pricing In One-way Car Sharing Networks:
A Distributional Fluid Approximation Approach
Ling Zhang, North Carolina State University, Raleigh, NC, United
States, Yunan Liu, Yang Liu, Shuangchi He
Balancing supply and demand across different areas is a critical issue in one-way
car sharing networks. We study dynamic pricing in order to maximize the profit
of a car sharing network. Since the stochastic network model is analytically
intractable, we propose a fluid approximation to represent the supply and
demand of vehicles. In contrast to conventional transportation fluid models that
assume deterministic processing times, general rental time distributions are built
into our fluid model. Moreover, our model allows for time-varying demand rates
and rental time distributions. Under this formulation, dynamic pricing is reduced
to a convex optimization problem that is efficiently solvable.
TB61
Cumberland 3- Omni
Online Delivery Routing
Sponsored: TSL, Freight Transportation & Logistics
Sponsored Session
Chair: Jan Fabian Ehmke, Freie Universität Berlin, Garystr. 21,
Berlin, 14195, Germany,
1 - Taking Advantage Of In-store Customers To Deliver Online Orders
Iman Dayarian, Georgia Institute of Technology,
765 Ferst Dr NW, Atlanta, GA, 30318, United States,
Martin W P Savelsbergh
Same-day delivery of online orders is becoming an indispensable service for large
retailers. We explore a novel environment in which in-store customers may take
over the task of delivering online orders on their way back home. Additionally, a
fleet of company-employed drivers is available to cover any unserved online
orders. This context represents a highly dynamic and stochastic environment for
which we explore and compare two rolling horizon approaches: one that ignores
any information about future arrivals of online orders and in-store customers,
and one that incorporates such information by means of sampled scenarios. Our
results demonstrate the superiority of scenario-based planning.
2 - An Online Cost Allocation Model For Horizontal Supply Chains
Han Zou, University of Southern California, Los Angeles, CA,
United States, Maged M Dessouky,
John Gunnar Carlsson
This research addresses the cost allocation problem in a real-time cost sharing
transportation system, which results from horizontal cooperation among multiple
suppliers. We formulate the cost allocation problem for the dynamic vehicle
routing environment, where only part of the customers are known in advance,
and the rest become known in real time. We propose an online cost-sharing
mechanism coupled with a look-ahead dynamic vehicle routing framework that
explicitly forecasts future customer requests.
3 - A Branch-and-price Approach For The Vehicle Routing Problem
With Roaming Delivery Locations
Gizem Ozbaygin, Bilkent University, Ankara, Turkey,
Martin W P Savelsbergh, Hande Yaman,
Oya Ekin Karasan
We study the vehicle routing problem with roaming delivery locations in which
each customer is associated with multiple locations and time windows. Exactly
one location per customer should be included in the delivery plan respecting the
time windows. We devise a branch-and-price algorithm to solve the problem and
perform a computational analysis.
4 - E-fulfillment For Attended Last-mile Delivery Services In
Metropolitan Areas
Jan Fabian Ehmke, Freie Universität Berlin, Garystr. 21,
Berlin, 14195, Germany, ,
Catherine Cleophas, Charlotte Köhler, Magdalena Lang
We consider service time windows as a scarce resource and combine concepts of
revenue management and vehicle routing to improve e-fulfillment of last-mile
delivery services. As the customer has to be present for attended deliveries such as
groceries, a service time window has to be agreed upon already when the order is
accepted. We will focus on the factors impacting the success for e-fulfillment in
metropolitan areas, considering uncertain demand and traffic conditions. To this
end, we analyse historical order data and extend time-dependent vehicle routing
techniques.
TB62
Cumberland 4- Omni
Data Mining and Optimization for Improved
Airport Operations
Sponsored: Aviation Applications
Sponsored Session
Chair: Heng Chen, University of Nebraska–Lincoln, Supply Chain
Management and Analytics, Lincoln, NE, 68588, United States,
1 - Airport Capacity Estimation For Decision Support
Sreeta Gorripaty, University of California Berkeley,
, Mark M Hansen
Capacity is a critical component of airport performance and air traffic decision-
making. Capacity of an airport is the throughput observed at sufficiently high
demand and is thus demand censored. The demand that is observed at the airport
is a result of strategic and tactical decisions made to avoid buildups of unmet
demand, thus making it challenging to estimate the capacity of an airport. We
demonstrate that Random Survival Forest (RSF) model can be used to capture the
censored nature of capacity data and model hourly capacity. The RSF capacity
model is further used in decision support algorithms to represent airport capacity.
2 - Parameter Fixing Method For Improving The Rate Of
Convergence Of A Hybrid Particle Swarm Optimization
Giuseppe Sirigu, Georgia Institute of Technology,
A new solution is proposed to perform just in time taxi operations using
autonomous electric towbarless tractors, thereby to minimize the overall cost and
the environmental impact of the ground operations. An algorithm for a tool that
provides conflict-free schedules for the tractor autopilots was developed, which is
based on a hybrid particle swarm optimization (HPSO), hybridized with a hill
climbing meta-heuristic. In order to improve the rate of convergence of the
algorithm, we developed a parameter fixing method.
3 - Machine Learning Techniques For Airport Passenger
Flow Management
Xiaojia Guo, University College London, London, United Kingdom,
, Yael S Grushka-Cockayne, Casey Lichtendahl,
Frederick Tasker, Neville Coss, Tom Garside, Bert De Reyck
Passengers missing their connection at an airport can have a major impact on
passenger satisfaction and airline delays. We develop a predictive model of
passengers’ connecting time using machine learning techniques, and provide both
point forecasts and probabilistic forecasts using historical and real-time data.
Based on these forecasts, we are developing a dynamic planning tool for London’s
Heathrow Airport to support airport operations.
TB61