2016 INFORMS Annual Meeting Program

102 / 561

INFORMS Nashville – 2016
102
SD25
110A-MCC
Managing Uncertainties in Projects
Invited: Project Management and Scheduling
Invited Session
Chair: Janne Kettunen, The George Washington University,
Washington, DC, United States,
1 - Zooming In On The Innovator’s Bias Within Organizations
Fabian Sting, Erasmus University Rotterdam, Rotterdam School of
Management, , Christoph Fuchs, Maik Schlickel
Firms in competitive industries strive for process innovations, and one source of
such ideas is the firm’s workforce. In the selection process, firms rely on input
from ideating employees - input that might contain systematic errors (biases)
and/or unsystematic errors (noise). We study such errors by analyzing the process
innovation ideas considered by an automotive manufacturer. Our data set is
unique in that it includes information on idea generation, employee evaluation,
standardized value calculation, selection, and implementation. Overall, our
findings contribute to a more differentiated yet theoretically coherent
understanding of the innovator’s bias in organizations.
2 - To Better Manage Risks In New Product Development Portfolio
Selection – Be Risk Neutral
Janne Kettunen, Assistant Professor, The George Washington
University, 2201 G Street, NW, Washington, DC, 20052,
United States, , Shivraj Kanungo
We investigate trade-offs between risk and return in multi-period new product
development (NPD) portfolio selection problems, where new development
projects become periodically available. Our analytical and computational results
show that, paradoxically, a risk-neutral NPD portfolio selection approach provides
higher return and lower risk than a risk-averse selection approach. This result can
explain why leading innovators tend to employ a risk-neutral NPD selection
approach. The risk of the NPD portfolio can be mitigated by (i) reviewing
portfolios more frequently and (ii) increasing the proportion of derivative
products instead of platform products.
3 - Project Portfolio Selection – A Behavioral Study
Sebastian Schiffels, Technical University of Munich, Munich,
80333, Germany, Thomas Fliedner,
Rainer Kolisch
Choosing the right set of projects is a key driver of success and failure in new
product development. We conducted experimental studies based on the knapsack
problem to address the question which decision rules individuals apply to select a
portfolio as well as how cognitive limitations influence their selection. Grounded
in portfolio selection practice, we investigate subjects’ adherence to four
heuristics. Decision making is partially explained by adherence to two simple
rules, but problem complexity limits the application of such rules as subjects apply
a local search. Furthermore, decision maker prefer projects with low risk resulting
in portfolios with few high risk high impact projects.
4 - Initiating Supplier New Product Development Projects:
A Behavioral Investigation
David Wuttke, EBS University, Wiesbaden, Germany,
, Karen Donohue, Enno Siemsen
Using a combination of analytical models and laboratory experiments, we study
the effectiveness of buyer contract mechanisms, including breach penalties and
profit sharing, on incentivizing product innovation at the supplier level. Our
results provide insight into how the mechanisms can be altered to better account
for supplier-specific behavior.
SD26
110B-MCC
Auctions and Trading Agents
Invited: Auctions
Invited Session
Chair: Wolfgang Ketter, Rotterdam School of Management, Rotterdam,
Netherlands,
1 - Using Optimal Grid Resources For Coordinating Electric
Vehicle Charging
Konstantina Valogianni, IE Business School, Madrid, Spain,
Alok Gupta, Wolfgang Ketter,
Soumya Sen, Eric F Van Heck
We propose a social welfare maximization mechanism to optimally schedule EV
charging, ensuring the lowest overall delay for the EV owners. At the same time,
our mechanism creates electricity peak demand reduction which is important for
improving sustainability in the grid. Our solution has lower computational
complexity, compared to state of the art mechanisms, making it easily applicable
to practice, where large numbers of EVs need to be charged. We prove the
theoretical optimal conditions that must hold in order to have maximum social
welfare in the grid. We validate our mechanism on real-world data and find both
peak demand and delay reduction.
2 - Truthful Approximation Mechanisms for Knapsack Bidders
Martin Bichler, Soeren Merting, Technische Universitat Munchen,
Munich, Germany.
In markets such as digital advertising markets, bidders want to maximize value for
impressions subject to a budget constraint. This type of utility function is typically
implemented in bidding agents, but it differs from quasilinear utility functions in
important ways. We refer to such bidders as knapsack bidders. We study the
offline mechanism design problem and analyze truthful approximation
mechanisms to maximize social welfare. Serial dictatorship mechanisms are
shown to be strategy-proof and Pareto-optimal, but they can have low welfare.
We propose a randomized mechanism with an approximation ratio of 4. Our
mechanism draws on a fractional deferred acceptance algorithm and randomized
rounding, and it illustrates how the relax-and-round principle can be
implemented in an important non-quasilinear environment.
3 - Modelling Electricity Balancing Market Prices And Premiums:
A Non-parametric Non-linear Approach
Ezgi Avci-Surucu, PhD Student, Rotterdam School of
Management, Rotterdam, Netherlands,
Wolfgang Ketter, Gerhard Wilhelm-Weber
In smart electricity markets, the increased penetration of renewable sources
reveals the need for decision support systems. For developing reasonable bidding
strategies, market participants need intelligent agents to make informed decisions
about the trade-off between sales in the day-ahead market or in the balancing
market. In this paper, by considering a detailed system-level data; firstly we
examine the market efficiency by fractal analysis to understand the level of price
predictability. Further, due the invalidity of normality and linearity assumptions,
we propose non-parametric non-linear models to provide strategic tools for policy
makers and market participants.
SD27
201A-MCC
Empirical Research in Finance and Operations
Sponsored: Manufacturing & Service Oper Mgmt
Sponsored Session
Chair: William Schmidt, Cornell University, United States,
1 - Optimal Timing Of Inventory Decisions Under Price Uncertainty
Nikolay Osadchiy, Emory University,
Vishal Gaur, Sridhar Seshadri, Marti Subrahmanyam
We study the problem of optimal inventory order timing when the selling price
and demand are random and their forecasts improve with time. We show that the
optimal timing of inventory ordering decision follows a simple threshold policy in
the price variable with a possible option of non-purchasing, and is independent of
the demand. Given this policy structure, we evaluate the benefits of timing
flexibility using the best pre-committed order timing policy as the benchmark.
2 - Wisdom Of Crowds: Forecasting Using Prediction Markets
Ruomeng Cui, Kelley School of Business, Indiana University,
Bloomington, IN, 47401, United States,
Achal Bassamboo, Antonio Moreno-Garcia
Prediction markets are virtual markets created to aggregate predictions from the
crowd. We examine data from a public prediction market and internal prediction
markets run at three corporations. We study the efficiency of these markets in
extracting information from participants. We show that the distribution forecasts,
such as sales and commodity prices predictions, generated by the crowds are
perfectly calibrated. In addition, we run a field experiment to study drivers of
forecast accuracy.
3 - Linking Operational Performance To Financial Distress In The
U.S. Airline Industry
Yasin Alan, Vanderbilt University, Nashville, TN, United States,
, Michael A Lapre
We study the impact of four areas of operational performance -revenue
management, operational efficiency, service quality and operational complexity-
on financial distress in the U.S. airline industry using quarterly data from 1988
through 2013. Our findings suggest that operational metrics convey useful
information regarding future financial distress even after controlling for financial
ratios that predict bankruptcies.
SD25