2016 INFORMS Annual Meeting Program

500 / 561

INFORMS Nashville – 2016
500
WE50
212-MCC
Opt, Nonlinear Programming II
Contributed Session
Chair: Carlos Deck, PhD Student, UC Berkeley, 4141 Etcheverry Hall,
University of California Berkeley, Berkeley, CA, 94720, United States,
1 - An Eigenvalue Decomposition Based Branch And Bound
Algorithm For Nonconvex Qcqp With Convex
Quadratic Constraints
Qingwei Jin, Zhejiang University, School of Management,
866 Yuhangtang Road, Hangzhou, 310058, China,
Cheng Lu, Zhibin Deng
In this paper, we propose a branch-and-bound algorithm for finding a global
optimal solution for a nonconvex quadratic program with convex quadratic
constraints (NQPCQC). We reformulate NQPCQC by adding some nonconvex
quadratic constraints induced by eigenvectors of negative eigenvalues associated
with the nonconvex quadratic objective function to Shor’s semidefinite
relaxation. Under the assumption of having a bounded feasible domain, these
nonconvex quadratic constraints can be further relaxed into linear ones to form a
special semidefinite programming relaxation. Then an efficient branch-and-bound
algorithm branching along the eigendirections of negative eigenvalues is
designed.
2 - Deployment Of Patrolling And Stationary Service Vehicles For
Freeway Incident Management
Mohsen Parsafard, University of South Florida, 14440 Hellenic
Drive, Apt 206, Tampa, FL, 33613, United States,
, Xiaopeng Li, Zhenyu Wang, Pei-Sung Lin
Freeway Service Patrols (FSP) are effective incident management systems to
alleviate the incident related congestion in urban areas. A FSP system divides a
network of freeways into a set of non-overlapping patrol beats and a number of
tow trucks are then assigned to these beats. This study investigates a FSP system
design problem. A mixed-integer nonlinear programming model is formulated for
designing the optimal beat configuration and service vehicle allocation to
maximally reduce delay, avoid secondary accident risks and mitigating fuel
consumptions and emissions. Also, this study considers not only traditional
patrolling vehicles or but also newly proposed stationary vehicles.
3 - Solutions Of Complementarity Problems Monotone With Respect
To Parameters
Vyacheslav V. Kalashnikov, Professor and Researcher, Tecnologico
de Monterrey, ITESM, Campus Monterrey, Ave. Eugenio Garza
Sada 2501 Su, Monterrey, 64849, Mexico,
Nataliya I. Kalashnykova, Mariel A. Leal-Coronado
In many applied problems (such as the elastohydrodynamic lubrication problem,
some economic equilibrium problems, etc.), one of the important questions is if
certain complementarity problem’s solution is monotone with respect to
parameters. Our paper investigates this question and provides several types of
sufficient conditions that guarantee such a monotonicity of the solutions to linear
and nonlinear complementarity problems with parameters. In the majority of
cases, it is required that the principal mapping of the complementarity problem be
monotone with respect to the decision variables and, vice versa, antitone with
respect to the parameters.
4 - Solving Lps Using Parametric Quadratic Programming
Carlos Deck, PhD Student, UC Berkeley, 4141 Etcheverry Hall,
University of California - Berkeley, Berkeley, CA, 94720,
United States,
Our approach for solving a standard LP with m constraints and n variables is
based on a parametric QP where we relax the inequality constraints of the
polyhedron of optimal solutions in primal-dual space. Using this approach, we
can solve the bad instances of the simplex method given by the deformed
products of Amenta and Ziegler (1999) in a polynomial number of pivots.
WE53
Music Row 1- Omni
DMA Text Mining
Contributed Session
Chair: Bill DeGray, IT Program Manager, Deloitte Services, LP,
Hermitage, TN, 37076, United States,
1 - Cognitive Analytics For Financial Compliance
Aditya Vempaty, IBM Research, Yorktown Heights, NY,
United States, , Elham Khabiri,
Swapna Buccapatnam, Matthew Riemer, Ta-Hsin Li,
Ashish Jagmohan
Financial services firms require compliance with a complex set of regulatory
requirements, laws, and regulators. In the face of increased regulatory
enforcement, firms have significantly increased compliance spend. A major
contributor to soaring compliance spend is the intensive expert manual
component of current compliance processes. We propose approaches for
automating key compliance processes using statistical machine learning and
natural language processing techniques. We present preliminary results
comparing performance against expert outcomes.
2 - On Determining The Viability Of Parametric Cost Forecasting Of
Agile Software Development Projects To Improve Project
Cost Forecasts
Bill DeGray, IT Program Manager, Deloitte Services, LP,
Hermitage, TN, 37076, United States,
This presentation describes methods being used in a four phased investigation to
determine whether parametric forecasting is a viable alternative to bottom-up
forecasting for software development projects. Phase 1 (completed) - design.
Phase 2 (completed) - determine if there is a relationship between the percentage
of the historical project costs incurred (the response variable) and the percentage
of the project schedule completed (the predictor variable). Phase 3 (in-progress) -
identifying the model(s) that best describe(s) the relationship. Phase 4 will focus
on development, testing, and piloting a parametric forecasting tool model based
on the best regression model(s).
WE54
Music Row 2- Omni
Value Co-creation
Sponsored: Service Science
Sponsored Session
Chair: Ralph D Badinelli, Professor, Virginia Tech, Blacksburg, VA,
24061, United States,
1 - Empirical Analysis Of Hospital Behaviors Resulting From Value
Based Purchasing
Gregory Heim, Mays Business School, Texas A & M University,
College Station, TX, 77845, United States,
We examine impacts of the federal Value Based Purchasing financial incentive
policy to determine whether this federal program has changed practices and
processes of healthcare providers in a manner that enhances process quality and
improves healthcare outcomes, thereby enhancing value for patients.
2 - Value Derivation In The Service Journey
Ralph Badinelli, Virginia Tech,
Value co-creation is accomplished in service systems through an adaptive journey
by service participants in a hypernetwork of service contexts. The viability of the
service ecosystem depends on the engagement decisions made by participants at
each context of the journey. Predicting and controlling this path is made complex
by the stigmergy enabled by social networking and other information sources. We
use fuzzy models of engagement decisions and examine the role of swarm
intelligence in determining the viability of service ecosystems. We justify a value
derivation principle in the place of value propositions and we derive patterns of
service journeys under different touchpoint strategies.
WE50