![Show Menu](styles/mobile-menu.png)
![Page Background](./../common/page-substrates/page0322.png)
INFORMS Nashville – 2016
320
TC52
214-MCC
Urban Transportation and Logistics in
Public Sector OR II
Sponsored: Public Sector OR
Sponsored Session
Chair: Sung Hoon Chung, Binghamton University, PO Box 6000,
Binghamton, NY, 13902, United States,
schung@binghamton.edu1 - A New Fast Algorithm For The Time-expanded Network Of
Dynamic Building Evacuation
Dong-jin Noh, Postech, Pohang, Korea, Republic of,
visionph@postech.ac.kr, Chang Hyup Oh, Young Myoung Ko,
Byung-In Kim
Time-expanded network models have been widely used in many evacuation
studies. Linear programming and heuristics algorithms have been applied for the
models. In this presentation, we propose a new fast exact algorithm for a time-
expanded network model arisen in a dynamic building evacuation. The proposed
algorithm takes advantage of the characteristic that there exist no cycle in the
time-expanded network model. Experimental results demonstrate the efficiency
of the proposed approach. The proposed algorithm can be applied for real-world
dynamic building evacuation problems.
2 - Human In The Loop Optimization For Recovery From
Extreme Events
Aybike Ulusan, Northeastern University, Boston, MA, 02115,
United States,
ulusan.a@husky.neu.edu, Ozlem Ergun
We consider the problem faced by contractors of collecting debris from a
transportation network to the disposal facilities in the aftermath of a disaster. The
problem has a multi-objective nature which embodies implementable division of
a service region among subcontractors such that the assigned workload among
different subcontractors is balanced and time to complete debris collection
operations is minimized. In this study, we explore the potential of having humans
collaborate with algorithms, and the use of game-based experiments to build a
decision support tool. We investigate how to exploit human input to achieve a
performance that can not be achieved by human or computer itself individually.
3 - Clearing Roads And Collecting Debris By Integrating Remote
Sensing Technique And Vrp
Eunsu Lee, New Jersey City University,
ELee3@njcu.eduEvery natural disaster generates large amount of debris on roads. Clearing the
debris from roads is crucial to recover mobility and accessibility to support
humanitarian aids and normal life. This study investigates an integrated approach
using remote sensing technology and vehicle routing problem to find, clear, and
dispose the items quickly and efficiently.
4 - Disaster Relief Routing under Uncertainty
Sung Hoon Chung, Binghamton University, 4400 Vestal Parkway
East, Binghamton, NY, 13902, United States,
schung@binghamton.edu, Yinglei Li
We propose a robust optimization approach for vehicle routing problems (VRPs)
under uncertainty for humanitarian logistics. In addition to classical cost-
minimizing and route-minimizing objectives of the VRPs, we employ alternative
objectives such as minimizing the average arrival time, the latest arrival time, and
the demand weighted arrival time, as it is critical for deliveries to be fast and fair
in routing for relief efforts. We show the use of the proposed approach using
benchmark problems and identify instances in which solutions of the VRP
variants are significantly different than ones of conventional VRPs.
TC53
Music Row 1- Omni
Management of Product and Service Innovation
Sponsored: Technology, Innovation Management &
Entrepreneurship
Sponsored Session
Chair: Juliana Hsuan, Professor, Copenhagen Business School, Solbjerg
Plads 3, B.5.27, Frederiksberg, DK-2000, Denmark,
jh.om@cbs.dk1 - Linking The Firm’s Internal Innovation Context With
Commercialization Choices
Lee Davis, Copenhagen Business School,
ld.ino@cbs.dk,Karin Hoisl, Jerome Davis
This paper investigates the linkages between the firm’s internal innovation
context and how the innovation is commercialized (by the firm itself or an
external third party). By analyzing these linkages at the level of the individual
innovation, we add to the literature on how firms profit from R&D. We base our
study on original survey data comprising 3,773 commercialized innovations from
23 countries in all major industries. We find that three aspects of the innovation
context - external knowledge inputs, encouragement of creativity, and high
autonomy - are positively related to external commercialization. Sufficient
resources are negatively related to external commercialization.
2 - Managing External Intellectual Capital In New Product
Development: The Case For Ontological Semantic Analysis Of
Patent Data
Charles Weber, Portland State University,
webercm@pdx.edu,Farshad Madani, Nitin Mayande
Historically, intellectual capital from outside the firm has been derived from
patent metadata. This paper presents a potentially much more effective approach,
which automatically analyzed the body text of patents.
3 - Logistics Service Performance In Nova Scotia: Facilitators,
Barriers, And Measurement
M. Ali Ulku, Rowe School of Business, Dalhousie University,
Halifax, NS, B3H3S7, Canada,
ulku@dal.ca,Horand I. Gassmann,
Michael Foster
Logistics plays a pivotal service role in efficient management of supply chains.
Building on the extant literature and company-survey results, we explore the
facilitators and barriers logistics companies face in the province of Nova Scotia,
Canada. We also propose key metrics for measuring the performance of regional
logistics services.
4 - Management Of Service And R&D Portfolios
Kai Basner, PhD Candidate, Copenhagen Business School, Solbjerg
Plads 3, Frederiksberg, DK-2000, Denmark,
kba.om@cbs.dk,Thomas Frandsen, Jawwad Raja, Juliana Hsuan
Managing technological innovation is critical to the continued success of
industrial companies, which in recent years have been observed to expand their
business models by complementing their products with services. For
manufacturers with a strong focus on product technology, we explore the
challenges of introducing service innovation in R&D portfolios.
TC54
Music Row 2- Omni
Service Science: Best Cluster Paper Presentation
Sponsored: Service Science
Sponsored Session
Chair: Paul Maglio, University of California, Merced, 5200 N Lake Rd,
Merced, CA, 95343, United States,
pmaglio@ucmerced.edu1 - Scheduling Promotion Vehicles To Boost Profits
Lennart Baardman, Massachusetts Institute of Technology,
Cambridge, MA, United States,
baardman@mit.edu,Maxime
Cohen, Georgia Perakis, Kiran Panchamgam, Segev Danny
With our collaborators from Oracle, we model how to schedule promotion
vehicles to maximize profits using ideas from the non-linear bipartite matching
problem. Promotion vehicles should be assigned to time periods, subject to
capacity constraints. We introduce a class of models for which the boost effects of
vehicles on demand are multiplicative. We show that the problem is
computationally intractable and develop a greedy method as well as a (1-epsilon)-
approximation using an IP of polynomial size. We analyze our methods as well as
validate them on actual data and finally, quantify their impact.
2 - Optimizing Precision In Machine Learning Models For Actionable
Predictions Of Revenue Change
Abhinav Maurya, Carnegie Mellon University, 5634 Stanton
Avenue, Apt 306, Pittsburgh, PA, 15206, United States,
ahmaurya@cmu.edu,Aly Megahed, Ray Strong,
Jeanette Blomberg, Alaa Elwany
Predicting changes in account revenues is of vital importance to a business in
order to take action on accounts predicted to shrink, and learn success stories of
offerings that led to revenue growth. However, the corresponding datasets are
often imbalanced, and so optimizing prediction accuracy, as the majority of
classifiers do, yields poor results in this case. We present a Gaussian Process-based
method that directly maximizes precision subject to a minimum recall level,
yielding actionable results without sacrificing much accuracy. Numerical
experiments show very promising results.
TC52