Informs 2015

75 / 552

INFORMS Philadelphia – 2015
73
SB23
23-Franklin 13, Marriott
Reflected Diffusions and Stochastic Networks
Sponsor: Applied Probability
Sponsored Session
Chair: Kavita Ramanan, Professor, Brown University, 182, George
Street, Box F, Providence, RI, 02912, United States of America,
Co-Chair: Mohammadreza Aghhajani, Brown University, Providence,
RI, United States of America,
1 - Reflected Brownian Motions in Stable Models for Large
Equity Markets
Ioannis Karatzas, Columbia University, Department of
Mathematics, New York, NY, 10027, United States of America,
We construct systems of Brownian particles interacting through their ranks. These
interactions give rise to invariant measures which are in broad agreement with
stability properties observed in large equity markets over long time-periods. The
models have connections with the analysis of Queueing Networks in heavy
traffic. Their hydrodynamic-limit behavior is governed by generalized porous
medium equations with convection. We survey briefly recent progress on some of
these fronts.
2 - Pathwise Differentiability of Semimartingale Reflected Brownian
Motions in Convex Polyhedrons
David Lipshutz, Postdoctoral Research Associate, Brown
University, 182 George Street Box F, Providence, RI, 02912,
United States of America, ,
Kavita Ramanan
We consider the pathwise differentiability of semimartingale reflected Brownian
motions (SRBMs) in convex polyhedrons with respect to input parameters that
determine the initial condition and drift. The derivatives are characterized as
solutions to time-dependent Skorokhod type problems associated with the
SRBMs. As applications, we characterize derivatives of stochastic flows for SRBMs
and compute the sensitivity of SRBMs to small perturbations of the drift.
3 - Existence and Nonexistence Results for Diffusion Limits of
Serve-the-Longest-Queue
Rami Atar, Technion, Department of Electrical Engineering,
Technion,, Haifa, Israel, Subhamay Saha
When serve-the-longest-queue is considered for a system with finite buffers that
are equal in size, diffusion limits of queue length exist in conventional heavy
traffic, while they do not always exist in the Halfin-Whitt regime. This
phenomenon will be described, along with implications on a game theoretic set
up of the network, in which customers are strategic.
SB24
24-Room 401, Marriott
Data Mining and Network Inference for Social and
Health Application I
Sponsor: Artificial Intelligence
Sponsored Session
Chair: Sung Won Han, New York University, 650 First Avenue,
New York, NY, United States of America,
Co-Chair: Chen Kan, University of South Florida, 4202 E. Fowler Ave.
ENB118, Tampa, FL, United States of America,
1 - The Evolution of User Roles in Online Health Communities
Xi Wang, The University of Iowa, S343 PBB, The University of
Iowa, Iowa City, IA, 52242, United States of America,
Zhiya Zuo, Kang Zhao
Online health communities (OHCs) have become a major source of social support
for people with health problems. By analyzing how users’ roles change over time,
we constructed transition graphs to illustrate evolution of users’ roles in an OHC.
It was revealed that the types social support received by a user may facilitate or
delay her role transitioning. Our research has implications to for OHC operators
to track users’ behaviors in order to manage an OHC.
2 - Extracting Biomedical Relationships from
Unstructured Documents
Adel Javanmard, Assistant Professor, Marshall School of Business,
University of Southern California, Los Angeles, CA,
United States of America,
The published biomedical research is growing at an unprecedented rate. On the
other hand, the information we need to extract for many research objectives is
becoming increasingly complex. This trend has driven the need for automated
curation of scientific literature. We consider the problem of extracting drug-gene
relationships from unstructured documents and propose a novel algorithm to this
end. Our algorithm is intimately related to new advances in clustering sparse
graphs.
3 - Network Models for Monitoring High-Dimensional Image Profiles
Chen Kan, University of South Florida, 4202 E. Fowler Ave.
ENB118, Tampa, FL, United States of America,
Hui Yang
Advanced sensing brings the proliferation of image profiles, which pose significant
challenges for process monitoring and control. However, traditional SPC is limited
in its ability to readily address complex structures of high-dimensional profiles.
This paper presents a novel dynamic network methodology for monitoring and
control of high-dimensional imaging streams. Experimental results show that the
proposed approach provides an effective online control chart for monitoring
image profiles.
SB25
25-Room 402, Marriott
Mobile and Social Data Analytics
Sponsor: Information Systems
Sponsored Session
Chair: Rajiv Garg, Assistant Professor, McCombs School of Business,
2110 Speedway, UT-Austin, Austin, TX, 78705,
United States of America,
1 - Mobile Big Data Analytics
Xueming Luo, Temple University, 1801 Liacouras Walk,
Philadelphia, PA, United States of America,
This presentation discusses how smart devices (smartphones, machine-to-
machine connected solutions, wearables, Internet-of-things technologies) and big
data affect advertising, promotions, marketing ROI, and omni-channel targeting
effectiveness. Over 3.6 billion people worldwide are deeply engaged with
smartphone devices. More than half of adult Americans rely on smartphones to
go about daily life.
2 - An Expository Study of Human Geographic Affinity Measurement
Based Mobility Network
Rong Duan, AT&T Labs, Research, 1 AT&T Way, Bedminster, NJ,
United States of America,
Geographic data play critical roles in different domains.Based on the nearly
complete coverage of mobile cell towers and the frequent activities of mobile
devices, we propose a Human Geographic Affinity (HGA) metric. HGA measures
the degree of recapture for a crowd in a specific geographic area during a time
period. The metric quantifies the stability of a crowd in a geographic area, and
categories areas by their dynamic of human mobility. This study is completed in-
line with privacy policy.
SB26
26-Room 403, Marriott
INFORMS Undergraduate Operations Research
Prize I
Cluster: INFORMS Undergraduate Operations Research Prize
Invited Session
Chair: Aurelie Thiele, Lehigh University, 200 W Packer Ave,
Bethlehem, PA, 18015, United States of America,
1 - A Composite Risk Measure Framework for Decision Making
under Uncertainty
Pengyu Qian, Columbia University, Columbia Business School c/o
PhD Office, 3022 Broadway,311 Uris Hall, New York, NY, 10027,
United States of America, , Zizhuo Wang,
Zaiwen Wen
In this talk, we present a unified framework for decision making under
uncertainty. Our framework is based on the composite of two risk measures
accounting for parametric (given distribution) and distributional uncertainty
respectively. The framework generalizes many existing models. We also propose
new models within this framework whose solutions have probabilistic guarantees
and are less conservative comparing to traditional models. Numerical experiments
demonstrate the strength of our models.
SB26