2015 Informs Annual Meeting

MB12

INFORMS Philadelphia – 2015

MB11 11-Franklin 1, Marriott

Advantage: Evidence from the Newspaper Industry Alessio Cozzolino, Assistant Professor in Strategy, University College Dublin, UCD Quinn School of Business, Dublin, Ireland, alessio.cozzolino@ucd.ie The dissertation explores the relationship between technological change and competitive advantage, studying the transformation of the Italian newspaper industry after the Internet (1995-2014). Using quantitative and qualitative methods, and building on a hand-collected longitudinal database and on a large set of interviews, Alessio theorizes and tests the consequences of a new type of technological change, one that destroys incumbents’ complementary assets while preserving their core know-how. 4 - Shifting LOCI of Innovation: A Study of Knowledge Boundaries, Identity and Innovation at NASA Hila Lifshitz-assaf, NYU, 100 Bleecker street, 13B, New York, NY, 10012, United States of America, h@nyu.edu This dissertation explores how the ability to innovate is being transformed by the Web and the information age,as well as the challenges and opportunities it entails.It is based on an in-depth longitudinal field study at NASA,exploring their experiment with online open innovation platforms and communities.I investigate the impact of using open innovation on the process of knowledge and innovation production,on R&D professionals,and its boundary conditions for successful problem solving. Sponsor: E-Business Sponsored Session Chair: Mingdi Xin, Assistant Professor, University of California, Irvine, Paul Merage School of Business, SB1-3423, Irvine, CA, 92697, United States of America, mingdi.xin@uci.edu 1 - Modeling the Dynamics of Network Technology Adoption and the Role of Converters Soumya Sen, University of Minnesota, Minneapolis, MN, United States of America, ssen@umn.edu, Youngmi Jin, Kartik Hosanagar, Roch Guerin We study the role of converters in the adoption of competing network technologies by heterogenous users. Converters can play an ambiguous role in the adoption process: they allow entrants to overcome the effect of incumbent’s user base but also introduce performance degradation and functionality limitations. Our analysis reveals a number of interesting and unexpected outcomes in this competition between network technologies. 2 - The Diffusion and Business Value of User Generated Content on Social Media: Evidence from Twitter Lanfei Shi, University of Maryland Smith School of Business, Van Social media platforms and user generated content (UGC) have become valuable marketing aids to firms; however, there is little systematic understanding of whether, and how, the diffusion of UGC through these platforms creates value for firms. Collaborating with one of the leading IT firms in the US, we examine the conditions under which the diffusion of UGC on Twitter adds value, with a specific focus on the role of content and user characteristics in creating value. 3 - Piracy and Information-goods Supply Chain Antino Kim, PhD Candidate, Foster School of Business, University of Washington, University of Washington, Seattle, WA, 98195ñ3200, United States of America, antino@uw.edu, Atanu Lahiri, Debabrata Dey In the presence of a retailer between the manufacturer and consumers of information goods, the legal channel faces two menaces; the internal issue of channel coordination and the external issue of piracy. We develop an economic model to study the interaction of the two. 4 - Issues in Supporting Older Versions of Software: A Game-theoretic Model Atanu Lahiri, University of Texas at Dallas, Jindal School of Management, Richardson, TX, 75080-3021, United States of America, atanu.lahiri@utdallas.edu, Debabrata Dey, Abhijeet Ghoshal A software manufacturer needs to stop supporting older versions of its product to encourage consumers to upgrade to the newest version. Consumers, however, can respond by holding out, to compel the manufacturer to do exactly the opposite, as it cannot really afford leaving too many nodes vulnerable in the user network. What should the manufacturer do then, and what are the welfare implications? Munching Hall, College Park, MD, 20742, United States of America, lanfeishi@rhsmith.umd.edu, Siva Viswanathan, Tianshu Sun MB10 10-Room 310, Marriott Economics of Digital Goods and Services

Discrete Decision Making and Computation Sponsor: Optimization/Integer and Discrete Optimization Sponsored Session Chair: Ruiwei Jiang, University of Michigan, 1205 Beal Ave., Ann Arbor, MI, 48109, United States of America, ruiwei@umich.edu 1 - A Comparison of SMIP Decomposition Methods Semih Atakan, PhD Student, University of Southern California, University Park Campus, Los Angeles, CA, 90089, United States of America, atakan@usc.edu, Suvrajeet Sen Many practitioners appear to use deterministic equivalent formulations (DEF), and off-the-shelf MIP solvers to address their SMIP problems. Since such approaches do not scale well, they have to remain satisfied with models that allow only a handful of scenarios. In contrast, SMIP decomposition is able to provide solutions to models whose DEF contain millions of mixed-integer variables in both stages. We compare results from a variety of such decomposition methods. 2 - Cut Generation Enhanced by Learning for Two-stage Stochastic Linear Programming Yiling Zhang, University of Michigan, 1205 Beal Avenue, Ann Arbor, MI, United States of America, zyiling@umich.edu, Siqian Shen We propose a new decomposition and cut generation paradigm for solving two- stage stochastic linear programs with complete recourse. Specifically, we “learn” from the Benders cuts and solution bounds from previous iterations (via Thomson Sampling or Imitation Learning approaches), to online select a subset of subproblems for deriving new cuts that could converge to optimum more quickly. Computational results are provided to demonstrate the efficacy of the approach. 3 - Forced and Natural Nestedness David Morton, Professor, Northwestern University, IEMS Department, Evanston, IL, 60208, United States of America, david.morton@northwestern.edu, Michael Nehme, Ali Koc We consider two combinatorial optimization problems in which we seek nested solutions. For the first, we formulate two types of two-stage stochastic integer programs to force nestedness. For the second, we maximize a supermodular gain function subject to a resource-availability constraint, and give conditions which ensure nested solutions at certain budget increments. A stochastic facility location problem, a graph clustering problem, and a chance-constrained program illustrate ideas. 4 - An Abstract Model for Branching and its Application to Mixed Integer Programming Pierre Le Bodic, Georgia Tech, Atlanta, GA, United States of America, lebodic@gatech.edu, George L. Nemhauser We present an abstraction of Mixed-Integer Programs (MIPs) to a simpler setting in which it is possible to analytically evaluate the dual bound improvement of choosing a given variable. We then discuss how the analytical results can be used to choose branching variables for MIPs, and we give experimental results that demonstrate the effectiveness of the method on MIPLIB instances where we achieve a 7% node improvement over the default rule of SCIP, a state-of-the-art MIP solver. MB12 12-Franklin 2, Marriott Recent Algorithmic Developments in Deterministic Global Optimization Sponsor: Optimization/Mixed Integer Nonlinear Optimization and Global Optimization Sponsored Session Chair: Yash Puranik, Carnegie Mellon University, Pittsburgh, PA, United States of America, ypp@andrew.cmu.edu 1 - Lagrangean Disjunctive Branch and Bound for Linear Generalized Disjunctive Programs Francisco Trespalacios, Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA, 15213, United States of America, ftrespal@andrew.cmu.edu, Ignacio E. Grossmann In this work we present a novel Lagrangean relaxation of the continuous relaxation of the (HR) reformulation of linear GDPs. This Lagrangean relaxation is simpler to solve than the (HR) and, even more relevant, its solution (a continuous LP) always yields 0-1 values for the binary variables. We present a disjunctive branch and bound for linear GDPs that exploits the proposed Lagrangean relaxation of the HR.

175