2015 Informs Annual Meeting

WB07

INFORMS Philadelphia – 2015

4 - Competitive Capacity Investment with Learning Youngsoo Kim, University of Illinois at Urbana-Champaign, Urbana, IL, ykim180@illinois.edu, Dharma Kwon, Kuno Huisman, Verena Hagspiel Two competing firms consider an investment with uncertain profitability. In particular, each firm needs to decide when to invest as well as how much to invest. Depending on the order of the investment, the leader and the follower naturally arise, and whoever becomes the follower can learn the leader’s performance and take that information into consideration. We investigate the equilibrium strategy in this game and study the impact of learning on the equilibrium. WB07 07-Room 307, Marriott Large-scale Portfolio Risk Cluster: Risk Management Invited Session Chair: Justin Sirignano, Stanford University, Huang Engineering Center, Stanford, CA, 93404, United States of America, jasirign@stanford.edu 1 - Large-Scale Loan Portfolio Selection Justin Sirignano, Stanford University, Huang Engineering Center, Stanford, CA, 93404, United States of America, jasirign@stanford.edu, Kay Giesecke, Gerry Tsoukalas The problem of optimally selecting a large portfolio of risky loans is a high- dimensional nonlinear integer program. Portfolios can range from thousands to even hundreds of thousands, making the problem extremely computationally challenging. Using weak convergence results, we develop an approximate optimization approach which is accurate and typically significantly faster than nonlinear integer program solvers. 2 - Modeling Large Portfolio Risks with Covariates Jianqing Fan, Princeton, Dept of Operations Res & Fin Eng, Princeton University, Princeton, NJ, 08544, United States of America, jqfan@princeton.edu, Weichen Wang, Yuan Liao We propose a flexible factor model for estimating large covariance matrices with covariates and introduce a Projected-PCA technique. We show that the unobserved latent factors can be more accurately estimated than the conventional. By using the newly proposed Projected-PCA, the rates of convergence of the smooth factor loading matrices are obtained, which are much faster than those of the conventional factor analysis. The proposed methods are illustrated by extensive numerical studies. 3 - Geometry of Defaults WB08 08-Room 308, Marriott Optimization Nonlinear Programming II Contributed Session Chair: Alex Feild, Human Systems Engineer, Naval Surface Warfare Center, 18444 Frontage Road, Ste 327, Dahlgren, VA, 22405, United States of America, alexander.feild1@navy.mil 1 - A Hub-and-Spoke Network Design with Nonlinear Penalty Function Ramez Kian, Bilkent University, Industrial engineering Department,, Bilkent University, Ankara, Turkey, ramezk@bilkent.edu.tr, Kamyar Kargar In this paper, a bi-objective uncapacitated hub location problem is addressed. The first objective is to minimize the traditional cost function and the second objective tries to maximize the amount of flow transported via the hub network. It corresponds to logistics companies which makes it impossible for them to route all the flow in the network. The mixed integer nonlinear model, the proposed solution approaches and numerical results are provided. 2 - Convex Relaxations in Gas Network Optimization Problems Pelin Cay, Lehigh University, 200 W Packer Ave, Bethlehem, PA, 18015, United States of America, pec212@lehigh.edu, Camilo Mancilla, Robert Storer, Luis Zuluaga Non-convexity, non-linearity and the need for a real time solutions are the main Richard Sowers, Professor, University of Illinois at Urbana- Champaign, Urbana, IL, 61801, United States of America, r-sowers@illinois.edu We develop a geometric picture of defaults in an Eisenberg-Noe setting. We adapt some recent work to generate global rankings from local (pairwise) comparisons based on shortfalls. This work is joint with Henry Schenck and Rui Song

challenges of gas network optimization problems. In this study, we show the benefits of using convex SOCP relaxations of the problem. This convexification provides lower bound information which can be useful in solving large scale real life problems more efficiently. 3 - Goal Reaching Kinematic Chain using a General Purpose Inverse Kinematic Approach Alex Feild, Human Systems Engineer, Naval Surface Warfare Center, 18444 Frontage Road, Ste 327, Dahlgren, VA, 22405, United States of America, alexander.feild1@navy.mil, Michael Hamilton, Patrick Mead The goal was to solve the problem of creating a general purpose model for partially analytical goal reaching kinematic chain. Our model is largely based on the Cyclic Coordinate Descent algorithm (CDD). We included a step in the cycle where the joint under evaluation was able to use its parent’s axial freedom to rotate its hinge axis prior to applying its own rotation. The model implementation took 56% fewer iteration and an estimated 27% fewer calculations to stabilize than a traditional CCD. Venture Capital Funding, Crowd Sourcing, New Product Development, and Supply Chain Transparency Sponsor: Technology, Innovation Management & Entrepreneurship Sponsored Session Chair: Zhili Tian, Assistant Professor, Florida International University, 11200 S.W. 8th Street, Miami, FL, United States of America, ztian@fiu.edu 1 - Too Successful to Innovate? Dynamic Entrepreneurial Reputation and Venture Capital This research integrates entrepreneurial experience, success, reputation and venture financing into a dynamic model which explores the execution of innovative projects under asymmetric information. We are interested in exploring the tension between the desire of the entrepreneur to exploit the current project opportunity and her understanding that the success or failure in this current project also impacts her future reputation. 2 - Hunters and Gatherers: Strategy Identification of the Leading Open Firms John Angelis, Elizabethtown College, 1 Alpha Drive, We analyzed 73 (of 100 original) firm open innovation efforts via their press releases. Using path and cluster analysis to analyze the coded data, we obtained statistically significant results. Our data shows at least two types of crowdsourcing: 1) Hunters, incumbent firms, learning open methodologies, that accelerate a single (often one time) experiment to implement crowd sourcing; or 2) Gatherers, often newcomers, born open, with less urgency that continuously interact with the crowd. 3 - Product Development by a Firm and its Supplier: Insights from a Problem-solving Approach Mohsen Jafari Songhori, Jsps Research Fellow, Tokyo Institute of Technology, J2 Bldg., Room 1704, 4259 Nagatsuta-cho,, Tokyo, 226-8502, Japan, mj2417@gmail.com, Takao Terano, Sunny S. Yang This study conceptualizes Product Development (PD) by a firm and a supplier as a problem solving process. The firms decompose the PD problem into sub-problems, and use different solution strategies (e.g. different design approaches with costs, quality). We investigate the firms’ problems (e.g. optimal design strategy) and their interactions (e.g. contract setting). 4 - An Empirically Grounded Model of Supply Chain Transparency Anton Shevchenko, York University, 4700 Keele St, Toronto, Canada, absh1986@gmail.com, Moren Lévesque, Mark Pagell, David Johnston Using multiple case studies our study explores how firms achieve the requisite transparency to manage their supply chains. We explore the barriers and enablers of transparency inherent in complex networks of buyer-supplier relationships and external stakeholder involvement before discussing firm strategies for managing escalating requirements for transparency. WB09 09-Room 309, Marriott Noam Shamir, Assistant Professor, Tel-Aviv University, Haim Levanon, Tel-Aviv, Israel, nshamir@post.tau.ac.il, David Zvilichovsky Elizabethtown, PA, United States of America, angelisj@etown.edu, John Ettlie, Joseph Miller

402