2016 INFORMS Annual Meeting Program

WC44

INFORMS Nashville – 2016

WC43 208A-MCC

2 - Optimal Timing Of Technology Adoption By Incumbents: War Of Attrition Versus Preemption Nick Huberts, Tilburg University, N.F.D.Huberts@uvt.nl I consider two incumbent firms with an option to adopt a differentiated technology. The firms decide upon both the investment moment and the investment size. I find that adoption kills the old technology only when innovation is radical. When the degree of innovation is small and when the products are not close substitutes a war of attrition arises. Otherwise the firms end up in a preemption equilibrium. When a second-mover advantage is present, firms either want to stay alone on the old market or want to set a larger capacity as Stackelberg follower. Market uncertainty increases the first-mover advantage and at the same time makes it more attractive for the endogenous follower to forego adoption. 3 - Disruptive Innovation In A Declining Market Kuno Huisman, Tilburg University, kuno.huisman@gmail.com The paper considers the problem of a firm operating in a declining market. The firm has an option to innovate and has to derive the right time to do so, if at all. We find that it can be optimal for the firm to innovate because of two reasons. The first reason is that a new technology is available with which the firm can achieve high profits. The second reason is that, due to demand saturation, profits of the established product have become so low that the firm will adopt a new technology even if the newest available innovation has not improved for some time. WC42 207D-MCC Behavioral Considerations in Pricing and Revenue Management Sponsored: Revenue Management & Pricing Sponsored Session Chair: Monire Jalili, The University of Oregon, 488 Lillis, Lundquist College of Business, Eugene, OR, 97403, United States, mjalili@uoregon.edu 1 - Dynamic Pricing And Learning In Prediction Markets Adam Schultz, University of Chicago-Booth School of Business, Chicago, IL, United States, adam.schultz@chicagobooth.edu, John Birge, N. Bora Keskin, Yifan Feng We consider a market maker who operates a prediction market for an event with an uncertain outcome (e.g., government election, sporting event, etc.) and must dynamically select a control (i.e., price) over time. We characterize the market maker’s optimal policy when the market includes only myopic agents and show how a myopic policy exhibits near-optimal performance. We also consider a market including a strategic agent who knows the event outcome (e.g., an insider trader) and demonstrate that the market maker’s policies are robust to the presence of a strategic agent in the market. 2 - Try Before You Buy Pricing. Should Rental Fees Apply To Purchases? Monire Jalili, The University of Oregon, mjalili@uoregon.edu, Michael Pangburn When a product has uncertain value or is used repeatedly over time, customers may opt to rent the product before purchasing. In some instances, sellers entice purchase conversions by offering part of the already-paid rental fee as a discount towards purchase. But, another common pricing tactic is for the seller to apply no such credit towards conversion to purchase. In this paper, we analyze the optimal pricing and discounting policy for a monopolist selling to a market of consumers facing uncertain product valuation, and derive the conditions under which a firm should optimally apply some of the rental price towards the product purchase. 3 - Analysts Decisions In Airline Revenue Management – An Experimental Study Claudia Schuetze, M.Sc., RWTH Aachen University, Aachen, Germany, claudia.schuetze@rwth-aachen.de, Catherine Cleophas Revenue management could, in theory, fully rely on automated systems to predict demand and optimize revenue. In practice, analysts play a crucial role for revenue management. They influence the system given additional information about market changes and the firm’s strategic objectives. We present a range of behavioural experiments to test how analyst decisions are affected by factors such as demand complexity and decision variables. Our analysis considers achieved revenue, learning effects, and decision biases. The aim is to prepare the ground for an improved decision support for revenue management analysts.

Information Elicitation Sponsored: Decision Analysis Sponsored Session Chair: Majid Karimi, Waterloo, ON, Canada, mk.majidkarimi@gmail.com 1 - Accept-reject Mechanisms For Team Formation

Yevgeniy Vorobeychik, Vanderbilt University, Nashville, TN, United States, eug.vorobey@gmail.com, Jian Lou, Martin van der Linden, Gregory Leo, Pranav Batra, Chen Hajaj, Myrna Wonders Team (coalition) formation has been studied from a number of perspective. However, treatment of this problem from the point of view of mechanism design has received relatively little attention, with few concrete and general mechanisms proposed. We describe and motivate a class of accept-reject mechanisms for this problem, and demonstrate their theoretical properties (both positive and negative). These mechanisms are computationally very challenging, and we describe several algorithmic approaches to these. 2 - Prediction Market Equilibria Via Substitutes And Complements Bo Waggoner, Harvard University, Computer Science, bwaggoner@fas.harvard.edu Based on joint work with Yiling Chen. I will propose definitions for when pieces of information, modeled as signals, can be considered substitutes or complements. We will see that substitutes (respectively, complements) characterize cases where prediction market participants rush to truthfully report (respectively, delay as long as possible). I will try to give a geometric picture for how probabilistic structure of signals and choice of scoring rule interact to produce substitutes or complements, and discuss implications for designing markets. 3 - Arbitrage-free Combinatorial Market Making Via Integer Programming Christian Kroer, Carnegie Mellon University, ckroer@cs.cmu.edu, Miroslav Dudík, Sébastien Lahaie, Sivaraman Balakrishnan We present a new combinatorial market maker that operates arbitrage-free combinatorial markets specified by integer programs. Although the problem of arbitrage-free pricing with bounded loss is #P-hard, we posit that the typical case might be amenable to modern integer programming (IP) solvers. At the crux of our method is the Frank-Wolfe algorithm which is used to implement a Bregman projection aligned with the market maker’s cost function, using an IP solver as an oracle. We demonstrate the tractability and improved accuracy of our approach on real-world prediction market data from combinatorial bets placed on the 2010 NCAA Men’s Division I Basketball Tournament. 4 - Making Science Of “Black Art”: Risk Bias In Market Scoring Rules Majid Karimi, University of Waterloo, Faculty of Engineering, mk.majidkarimi@gmail.com, Stanko Dimitrov We study market scoring rules (MSRs) prediction markets (PMs) in the presence of risk averse or risk seeking agents. We show that agents’ submitted reports always deviate from their beliefs. This means, in most cases it is impossible for a MSR PM to elicit an agent’s belief. We introduce a measure to calculate the deviation between an agent’s report, and her personal belief. We find that the deviation of a MSR PM is related to the amount of market depth provided by the MSR’s cost-function PM. We use the relation between deviation and market depth to present the first systematic approach to determine the optimal amount of market depth, an activity that has been described as “black art” in the literature. WC44 208B-MCC Strategic Management Decision Making Sponsored: Decision Analysis Sponsored Session Chair: Dharma Kwon, University of Illinois at U-C, Champaign, IL, United States, dhkwon@illinois.edu 1 - Dynamic Sourcing Decisions In Presence Of Technology Spillover Risks Yunke Mai, Duke University, yunke.mai@duke.edu, Sasa Pekec We study optimal dynamic sourcing decisions of a serial innovator. There are two types of manufacturers: competitive ones who might pose technology spillover risks, and non-competitive ones. Manufacturers’ production capabilities are uncertain, impacting success of innovations. Single period contracts allow learning about the uncertainty by observing the production outcome. Long-term contracts lock the innovator with one manufacturer but guarantee a low wholesale price. We describe optimal strategies and show that contracting with a competitive manufacturer could be attractive as it allows for sharing the innovation risk in exchange for the technology spillover exposure.

441