Informs Annual Meeting 2017

TA28

INFORMS Houston – 2017

3 - Evaluating the Potential of Additive Manufacturing for Spare Parts Supply Rob Basten, Eindhoven University of Technology, P.O. Box 513, IE&IS.- OPAC, Eindhoven, 5600 MB, Netherlands, r.basten@utwente.nl, Bram Westerweel, Geert-Jan Van Houtum We investigate the decision of whether or not to replace a regularly produced spare part with an alternative produced via additive manufacturing (AM), also called 3D printing. Such a transition requires a one-time investment and changes characteristics such as procurement lead time, production costs and component reliability. We consider the entire product lifecycle and take into account performance benefits and design, inventory, maintenance and downtime costs. We derive analytical properties of the required reliability and production costs of the AM part such that its lifecycle costs break even with that of its regular counterpart. 4 - 3D Printing vs. Traditional Flexible Technology: Implications for Operations Strategy Duo Shi, Olin Business School, Washington University-St Louis, K.H.401, Olin Business School, 1 Brookings Drive, St. Louis, MO, 63130, United States, dshi@wustl.edu, Lingxiu Dong, Fuqiang Zhang We study a firm’s operational strategy under two types of flexible production technologies: the traditional flexible technology and 3D printing. Under the traditional flexible technology, capacity becomes more expensive as it handles more product variants; under 3D printing, however, capacity cost is independent of the number of product variants processed. The firm possesses a dedicated technology and may additionally adopt one or both of the two types of flexible technologies. Our results demonstrate that the rising 3D printing has significantly different implications for firms’ assortment and production strategy than the traditional flexible technology. 350C Network Analysis with Social Media Data Invited: Social Media Analytics Invited Session Chair: Alexander Nikolaev, University at Buffalo, Buffalo, NY, 14260- 2050, United States, anikolae@buffalo.edu 1 - Explaining the Emergence of Power Laws in Online Social Networks We present a method for parameterizing a network formation model that explains how the information about “socially contagious” individuals/items propagates into the masses, and the power law in degree emerges as these individuals/items reach new followers. We report the inferred model parameter values for several real-world social media datasets. 2 - Clustering of Ordered Restaurant Rankings on Yelp.com Alexander Semenov, University of Jyväskylä, Jyväskylä, Finland, alexander.v.semenov@jyu.fi, Alexander Nikolaev The ease of posting anonymous reviews online has created many opportunities for people to express their opinions about products and services in social media sites. The difference in opinions may or may not imply malicious behavior; in any case, identifying multiple groups of users, sharing different distinct views on how the evaluated items fair against each other, may help us better understand the diversity in tastes. We present a method to identify distinct user groups based on how they rank restaurants on Yelp.com, and interpret the results. 3 - Network Analysis of Social Media Portal VK.com Vladimir Boginski, University of Central Florida, Orlando, FL, 32541, United States, Vladimir.Boginski@ucf.edu, Alexander Nikolaev, Alexander Veremyev, Eduardo Pasiliao, Alexander Semenov The “post-Soviet space” includes a group of countries with a substantial fraction of the world’s population and spans multiple time zones and cultures, creating a unique geopolitical mixture. We present the study of the social network of VK.com (also known as VKontakte), which is the largest European (mainly Eastern-European) social portal, with over 300 million nodes. We analyze several aspects of the VK network structure, including its global connectivity patterns and community composition and draw conclusions about the “social media landscape” of the post-Soviet space. TA27 Rakesh Nagi, SUNY- Buffalo, Buffalo, NY, United States, nagi@illinois.edu, Sushant Khopkar, Alexander Nikolaev

4 - Engagement-based Predictive Analyses of the Growth of Online Health Rahul Gopalsamy, University at Buffalo, 409 Bell Hall, Buffalo, NY, 14260-2050, United States, rahulgop@buffalo.edu, Alexander Nikolaev, Scott McIntosh, Alexander Semenov, Eduardo Pasiliao The research on nurturing the growth of online communities posits that it is in part attributed to the network externality effects. This study uses the data of a large online portal to show that user engagement - measurable using the recently introduced metric “engagement capacity” - has a positive causal effect on online health community growth, opening the door to exploiting network externality effects in calculated ways.

TA28

350D Behavioral Models in Auctions

Invited: Auctions Invited Session

Chair: Kyle Woodward, University of North Carolina at Chapel Hill, Chapel Hill, NC, 27599-3305, United States, kyle.woodward@unc.edu 1 - Auctions with Max-min Players Kyle Woodward, University of North Carolina at Chapel Hill, Department of Economics, Gardner Hall, CB #3305, Chapel Hill, NC, 27599-3305, United States, kyle.woodward@unc.edu, Marek Pycia Multi-unit auctions are commonly used to allocate government debt and electricity generation, but little is known about equilibrium strategies under private information. We show that when bidders have max-min beliefs over opponents’ information, equilibrium is analytically tractable; our results apply to both private- and interdependent-values models. We define a natural equilibrium selection, upside dominance, implying equilibrium uniqueness; we show upside dominance to be equivalent to limiting forms of risk and ambiguity aversion. In upside-dominant equilibrium pay-as-bid auctions revenue-dominate uniform- price auctions, offering novel evidence favoring pay-as-bid auctions. 2 - Declining Prices Across Second-price Reverse Auction Formats Marion Ott, RWTH Aachen University, Templergraben 64/V, Aachen, 52062, Germany, marion.ott@rwth-aachen.de, Karl-Martin Ehrhart In four experimental reverse second-price auction formats for procuring a good for which bidders have uncertain private values, we find declining auction prices across the four formats. Pseudo-endowment effect arguments on psychological ownership, detachment, and willingness-to-pay/willingness-to-accept disparity predict the order of the declining prices. We conclude that auctions that foster detachment from an item may lead to lower payments. 3 - Empirical Relevance of Ambiguity in First-price Auctions Gaurab Aryal, University of Virginia, VA, United States, aryalg@virginia.edu We study the identification and estimation of the first-price auction model with IPV when risk averse bidders face ambiguity about the value distribution and have maxmin expected utility. We use variation in the number of bidders to nonparametrically identify the true value distribution and the lower envelope in the bidders’ set of priors. We propose a flexible Bayesian estimation method based on Bernstein polynomials. MC experiments show that our method estimates parameters precisely and chooses the reserve prices with (nearly) maximal revenues. We apply our method to U.S. timber auctions and find evidence of ambiguity in auctions from the Pacific Northwest region. 4 - Threshold Problems in Ascending Combinatorial Auctions Bart Vangerven, KU. Leuven, Naamsestraat 69, Leuven, 3000, Belgium, bart.vangerven@kuleuven.be, Dries Goossens, Frits C.Spieksma We experimentally study bidding behavior in ascending combinatorial auctions with threshold problems, using human bidders. We vary feedback from very basic information about provisional allocations, to more advanced concepts as winning/deadness levels, and even coalitional feedback, aimed at helping bidders to overcome threshold problems. The main question we address is the following: `Does additional feedback help bidders overcome threshold problems in ascending combinatorial auctions?’ We test this in different auction environments, investigating the effects on economic efficiency, auction revenues, duration, etc.

279