2015 Informs Annual Meeting

MD12

INFORMS Philadelphia – 2015

MD12 12-Franklin 2, Marriott

2 - Has Production Interdependence Been Changed with Information Technology?

Fengmei Gong, Assistant Professor Of Information Technology, University of La Verne, La Verne, CA, 91750, United States of America, fgong@laverne.edu, Barrie R. Nault, Zhuo (june) Cheng Industries have become increasingly integrated with their suppliers’ business processes such as purchasing and Just-in-time (JIT) production; however, whether industries in a supply chain have become more interdependent remains an open question. We examine the impact of an industry’s IT investment on its production interdependence with upstream suppliers, where we measure interdependence as direct backward linkage (DBL). 3 - New Platform Announcement Strateges: A Duopoly of Two-sided Platforms Rajiv Mukherjee, Assistant Professor, Southern Methodist We study a duopoly where two firms that are horizontally differentiated in their two sided platform offerings evaluate their release strategies for a new version. The firms chose between two release strategies: I) Formal release whereby the firms commit to their future offering, II) Informal release whereby the firms employ rumor or other seeding mechanisms to announce to the market. 4 - Social Interactions and Product Sales in Social Shopping: An Experimental Approach Annibal Sodero, Assistant Professor, University of Arkansas, Sam M. Walton College of Business, Fayetteville, AR, 72701, United States of America, ASodero@walton.uark.edu, Elliot Rabinovich, Bin Gu Social shopping revolves around deeply discounted deals that are offered for a limited time through social networking websites. In this study, we investigate the effect of social interactions on product sales and the contingencies surrounding the interactions. Using an experimental approach, we investigate five social interaction mechanisms and find that three mechanisms act in tandem to accelerate a deal’s demand: opinion leadership, network integration, and boundary spanning of early buyers. MD11 11-Franklin 1, Marriott Convex Relaxations for Structured Integer Programs Sponsor: Optimization/Integer and Discrete Optimization Sponsored Session Chair: Akshay Gupte, Clemson University, Dept of Mathematical Sciences, Clemson, SC, 29634, United States of America, agupte@clemson.edu 1 - On The Polyhedral Structure of a Multi-capacity Mixing Set Ayse Arslan, PhD Student, University of Florida, Weil 413, In this talk, we study the polyhedral structure of a multi-capacity mixing set. This set arises as part of the formulation of production planning and logistics problems. We derive two families of facet-defining inequalities for the set under consideration by lifting mixing inequalities. We discuss the properties of the associated lifting function and show that lifting can be performed efficiently. We thereby strictly generalize earlier results of Marchand and Wolsey [1998]. 2 - Sparse Principal Component Analysis (SPCA) via Convexification Jinhak Kim, Purdue University, 610 Purdue Mall, West Lafayette, IN, 47906, United States of America, kim598@purdue.edu, Mohit Tawarmalani, Jean-philippe P. Richard We characterize the convex hull of the feasible set of SPCA. The convex hull is described in a lifted space by dualizing the separation problem. The convex hull can be reformulated in terms of majorization inequalities. This interpretation allows us to express each point in the convex hull as a convex combination of points that satisfy the cardinality constraint. We propose an SDP relaxation in the lifted space which is stronger than that of d’Aspremont et al (2007). 3 - A Bilevel Programming Problem Occurring in Smart Grids Leo Liberti, CNRS & Ecole Polytechnique, LIX Ecole Polytechnique, Palaiseau, France, liberti@lix.polytechnique.fr, Sonia Toubaline, Pierre-louis Poirion, Claudia D’Ambrosio A key property to define a power grid “smart” is its real-time, fine-grained monitoring capabilities. For this reason, a variety of monitoring equipment must be installed on the grid. We look at the problem of fully monitoring a power grid by means of Phasor Measurement Units (PMUs), which is a graph covering problem with some equipment-specific constraints. We show that, surprisingly, a bilevel formulation turns out to provide the most efficient algorithm. University, Dallas, TX, United States of America, rmukherjee@mail.smu.edu, Ramnath Chellappa Gainesville, FL, 32611, United States of America, arslan.aysenur@gmail.com, Jean-philippe P Richard, Yongpei Guan

Surrogate-Based and Derivative-Free Optimization II Sponsor: Optimization/Mixed Integer Nonlinear Optimization and

Global Optimization Sponsored Session

Chair: Rommel Regis, Saint Joseph’s University, Mathematics Department, 5600 City Avenue, Philadelphia, PA, 19131, United States of America, rregis@sju.edu 1 - A DFO-based Approach to Computer-aided Mixture Design Nick Austin, Graduate Student, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, PA, 15213, United States of America, ndaustin@andrew.cmu.edu, Nikolaos Sahinidis, Daniel Trahan Computer-Aided Mixture Design (CAMxD) relies on complex physicochemical simulation models to design a blend of compounds. We present a novel approach to CAMxD that relies on the use of derivative-free optimization (DFO). We present comparative results from the application of 27 DFO solvers to this challenging problem. 2 - Surrogate-based Optimization for Oral Solid Drug Product Manufacturing Zilong Wang, Graduate Research Assistant, Rutgers University, 98 Brett Rd, Chemical and Biochemical Engineering, Piscataway, NJ, 08854, United States of America, wzlpublic@gmail.com, Surrogate-based optimization is used to solve computationally expensive simulation models and to optimize functions when the model is not available. However the applicability of such methods can be limited due to the high dimensionality of problem variables. In this presentation we focus on solving high-dimensional design problems in pharmaceutical manufacturing using RBF- based surrogate modeling strategies. Case studies will be used to illustrate the applicability of the proposed approaches. 3 - Applied Results from the Techno-economic Optimization of a High-flux Solar Thermal Receiver Michael Wagner, Mechanical Engineer, National Renewable Energy Lab, 15013 Denver West Parkway, Golden, CO, 80401, United States of America, Michael.Wagner@nrel.gov, Alexandra Newman, Robert Braun We optimize a novel concentrating solar power tower receiver technology by choosing the geometry and optical design. We use computationally expensive engineering models to generate surrogates that represent the objective function, which accounts for revenue as a function both of the design of the system and of the annual plant electricity production. Nonlinear constraints are incorporated via Lagrangian terms. We present results that guide the applied technology configuration. 4 - Applications of Surrogate-based Optimization Cameron Turner, Associate Professor Of Mechanical Engineering, Colorado School of Mines, 1500 Illinois St., Department of Many engineering design problems are characterized by nonlinear behaviors, mixed discrete-continuous variables, multiple objective functions, & uncertain or limited precision data about the problem. What data that exists is often derived from empirical measurements, experimental studies, or models & simulations; each with errors, limited precision & data collection costs. We focus on the use of the techniques, tradeoffs and decisions necessary to employ surrogates in optimization. M. Sebastian Escotetespinoza, Ravendra Singh, Fernando J. Muzzio, Marianthi Ierapetritou Mechanical Engineering, Golden, CO, 80401, United States of America, cturner@mines.edu

235