2015 Informs Annual Meeting

WD39

INFORMS Philadelphia – 2015

2 - What Can Merger and Acquisition (M&A) do for Healthcare Services? Yujiun Tsai, Texas A&M University, 3131 TAMU, College Station, TX, United States of America, yjt2009@tamu.edu We would like to explore the rationalization of hospitals using M&A to differentiate themselves through services, quality measures, and customer satisfaction. 3 - The Impact of Knowledge Sharing on Healthcare Risk Management Performance Mecit Can Emre Simsekler, Research Associate, UCL School of Considering two key components of knowledge sharing among healthcare personnel, (i) codified in the form of written documents and (ii) tacit with behaviors and daily practices, we examine how knowledge sharing capabilities in healthcare settings translate into risk management performance. We employ a unique dataset from the NHS acute trusts in England to investigate our hypotheses. 4 - Infectious Disease Outbreak Response Strategies for Ebola Neil Desnoyers, Instructor, Saint Joseph’s University, 133 Green Valley Rd, Upper Darby, PA, 19082, United States of America, ntdesnoyers@gmail.com High viral-load Ebola patients cause an outsize proportion of all transmissions. Outbreak response strategy should include the construction and/or use of permanent, temporary, and/or mobile healthcare facilities, simultaneously. I investigate the use of information on the requirements of high vs. low viral-load Ebola patients. I discuss optimal healthcare facility use to inform Ebola outbreak response and facility deployment strategies, thereby achieving maximum benefit at minimum cost. WD38 38-Room 415, Marriott Optimization Combinatorial III Contributed Session Chair: Pascal Rebreyend, Senior Lecturer, Högksolan Dalarna, Rödavagen, Falun, 79188, Sweden, prb@du.se 1 - The J-set and Successive Mccormick Relaxations for Polynomial Programming Problems Evrim Dalkiran, Wayne State University, 4815 4th St. MEB # 2149, Detroit, MI, 48202, United States of America, evrimd@wayne.edu We analyze the relative strength and tractability of the two linear programming relaxations obtained by the J-set of constraints constructed for the original polynomial formulations and the McCormick relaxations constructed for equivalent quadratic formulations via successive quadrification scheme. We propose a hybrid algorithm that judiciously selects among the J-set relaxation and the McCormick inequalities based on the problem’s structural characteristics. 2 - Symmetry: What LP Can Learn from MIP Management, UCL, Gower Street, London, WC1E 6BT, United Kingdom, e.simsekler@ucl.ac.uk, Bilal Gokpinar

experimental context. Tested methods simulated annealing, volume algorithm and Cplex. Our new hybrid genetic approach outperforms other existing approaches on large instances. 5 - Conic Least Squares Problem Yu Xia, Assistant Professor, Lakehead University, 955 Oliver Rd, Thunder Bay, ON, P7B 5E1, Canada, yxia@lakeheadu.ca I give two reformulations of the dual of the constrained least squares problem over convex cones. The conic least squares problem is then solved by applying modified Nesterov’s excessive gap method or Nesterov’s smooth method. Numerical experiments comparing this approach with interior-point method based state-of-art software are given.

WD39 39-Room 100, CC Marketing/Operations Management Cluster: Operations/Marketing Interface Invited Session Chair: Rachel Chen, Univeristy of California, Davis, CA,

United States of America, Associate Professor, rachen@ucdavis.edu Co-Chair: Cuihong Li, University of Connecticut, 2100 Hillside Rd, Storrs, CT, United States of America, Cuihong.Li@business.uconn.edu 1 - Retailer Adoption of Innovative Products

Jane Gu, Assistant Professor, University of Connecticut, 2100 Hillside Road, Storrs, CT, United States of America, jane.gu@business.uconn.edu, Yunchuan Liu

We consider a research context where innovators with new product development expertise lack direct-selling capabilities and big retailers control the access to the consumer market. We examine how a retailer decides whether to carry an innovator’s product based on its private information regarding the market acceptance to the innovative product. Our investigation reveals the interesting impact of vertical and the horizontal channel structures on the retailer’s innovation adoption decision. 2 - Product Line Design: Variety and Responsiveness Cuihong Li, University of Connecticut, 2100 Hillside Rd, A larger product variety allows a firm to better satisfy the needs of heterogeneous customers, but, in a make-to-order environment, it increases the job completion time, leading to longer waiting of customers. We study the trade-off between product variety and responsiveness to align product line design and operations system design. 3 - Why and Where to Have Outlet Stores? Shuya Yin, University of California, Irvine, Merage School of Storrs, CT, 06269, United States of America, Cuihong.Li@business.uconn.edu, Laurens Debo Outlet stores have been both complementary to and competing with the main stores. In this project, our goal is to understand the tradeoffs involved in offering outlet stores. In particular, we study how much product differentiation and physical distance should be kept between the main and outlet stores. 4 - Open Or Closed? Technology Strategy, Supplier Investment, and Competition Bin Hu, Assistant Professor, UNC Kenan-Flagler Business School, CB#3490 McColl Bldg, University of North Carolina, Chapel Hill, NC, 27519, United States of America, Bin_Hu@kenan- flagler.unc.edu, Ming Hu, Yi Yang We analyze a model where each of two manufacturers decides whether to open its technology to the competitor, and a supplier subsequently decides what technologies to invest in. We find that open technology incentivizes supplier investment, and also leads to a technology-risk-pooling benefit. The manufacturers may also be faced with the prisoner’s dilemma. Finally, we show that manufacturers may close their technologies to force the supplier to make a technology investment. 5 - Money-back Guarantees When Physical and On-line Retailers Compete Hang Ren, University College London, London, United Kingdom, hang.ren.13@ucl.ac.uk, Tingliang Huang, Chris Tang, Ying-ju Chen We study the pricing and product return policies when physical and on-line stores compete. We find that the on-line store offers money-back guarantees when its salvage advantage outweighs total return hassle. Interestingly, better service quality may hurt the on-line store. When consumers can showroom, i.e. buying online after trying the product offline, we show that the on-line store should offer hassle-free money-back guarantees. Business, Irvine, CA, United States of America, shuya.yin@uci.edu, Jiarui Bai, Haresh Gurnani

Roland Wunderling, IBM, Annenstrasse 9, Graz, Austria, roland.wunderling@at.ibm.com, Jean-francois Puget

Symmetry has long been exploited in the solution of mixed integer programs. While LP does not suffer from the same combinatorial explosion of the search space due to symmetry as MIP does, symmetries can be identified and exploited

for LP as well. We will evaluate the effect of doing so. 3 - Profit-oriented Ring Arborescence Problems

Alessandro Hill, Hamburg University of Technology, Schwarzenbergstrasse 95 D, Hamburg, Germany, alessandro.hill@tuhh.de, Roberto Baldacci, Edna Hoshino

In this work we study three new problems in extended network design. Two types of customer nodes and Steiner nodes can be used in a two-level network. Type two customers have to be in circuits that intersect in a depot. Type one customers may also be used in arborescences that extend this ring core. Objectives take into account the arc costs, customer-dependent profits or both. We present MIP models, valid inequalities, corresponding exact algorithms, heuristics and computational results. 4 - Testing Algorithm for Large P-median Problems in Heterogenous Road Networks Pascal Rebreyend, Senior Lecturer, Hügksolan Dalarna, Rüdavagen, Falun, 79188, Sweden, prb@du.se, Laurent Lemarchand This paper presents and compares different algorithms on large scale p-median problems, up to 2000 candidate nodes. Our main focus is instances where the demand is asymmetric distributed. We use as real data the Swedish road network including distances and as demand points the location of Swedish citizens as our

465