2016 INFORMS Annual Meeting Program

222 / 561

INFORMS Nashville – 2016
222
2 - Production Of Schooling
John Ruggiero, University of Dayton,
In this paper we analyze the production of schooling using stochastic DEA. Using
Banker’s stochastic DEA model, we estimate frontier production allowing for
measurement error while controlling for the socioeconomic environment. We
derive useful policy measures related to costs to help guide school funding.
3 - Insights From Machine Learning For Evaluating Production
Function Estimators On Manufacturing Survey Data
Andrew Johnson, Texas A & M University-College Station,
, José Luis Preciado Arreola
Organizations like census bureaus rely on non-exhaustive surveys to estimate
industry population-level production functions. In this paper we propose selecting
an estimator based on a weighting of its in-sample and predictive performance on
actual application datasets. We compare Cobb-Douglas functional assumptions to
nonparametric shape constrained estimators. For simulated data, we find that our
proposed estimator has the lowest weighted errors. For actual data, specifically
the 2010 Chilean Annual National Industrial Survey, a Cobb-Douglas specification
describes at least 90% as much variance as the best alternative estimators in
practically all cases considered.
4 - Endogenous Environmental Variables In Stochastic
Frontier Models
Artem Prokhorov, University of Sydney Business School,
Sydney, Australia, ,
Christine Amsler, Peter Schmidt
We consider an SFA model with errors e=v+uoexp(q’d), where v is normally
distributed noise, uo generates technical inefficiency and q are the environmental
variables that influence the level of technical inefficiency. In a previous paper we
looked at the case that the input variables are correlated with v and u but q are
exogenous. Here we allow q to be endogenous in the sense that they are not
independent of v and/or uo. We consider estimation by IV and by MLE. MLE
requires specification of reduced form equations. The case that q and uo are
dependent is difficult because we need to assume a copula and so there are
computational issues.
MD41
207C-MCC
Quantitative Methods for Finance and Energy
Sponsored: Financial Services
Sponsored Session
Chair: Mendoza Rafael,
1 - Modeling Dependent Outages Of Electricity Generators
Vishwakant Malladi, McCombs School of Business,
We present a framework where the electricity plants in a region are modeled as
subordinated Markov Chains. We also develop a factor model for Markov chain
generators to separate both the idiosyncratic and correlated behavior of the
plants. Calibration shows that supply curves are bent resulting in lower
generation capacity available at higher reliability levels.
2 - Dynamic Mean-variance Under Predictable Criteria
Xiao Han, University of Texas-Austin,
While mean-variance analysis has been widely adopted by investment
professional as a basic asset allocation tool, its dynamic counterpart was rarely
studied in academic literature. We aim to fill this gap by proposing a type of
mean-variance criteria that is predictable, self-generating and consistent over
time. The framework partially resolves the dilemma when searching for dynamic
optimal portfolio strategies while facing uncertainties about model parameters
and investment horizons. As an example, we solve the forward mean-variance
problem when the investor only has partial information regarding the dynamics
of asset returns.
3 - Static Hedging And Pricing Under The JDCEV Model Via
Integral Equations
Dong-Young Lim, Korea Advanced Institute of Science &
Technology, DaeJeon, Korea, Republic of, ,
Kyoung-Kuk Kim
We provide a systematic approach to construct an exact static hedge for exotic
options under the JDCEV model, using the theory of integral equations. We show
the existence and uniqueness of an exact static hedging portfolio which consists
of continuum of vanilla or binary options. Under suitable conditions, such a
hedging portfolio can be explicitly found in terms of generalized hypergeometric
functions. Also, we work on constructing a robust static hedge with finitely many
hedging instruments, together with an efficient method of evaluating hedge
errors. The effectiveness of the proposed method is demonstrated by several
examples, including double barrier options and step up(down) options.
MD42
207D-MCC
Pricing and Information Provision of Services
Sponsored: Revenue Management & Pricing
Sponsored Session
Chair: Gad Allon, Northwestern University, Evanston, IL,
United States,
1 - Managing Customer Expectations And Priorities With Delay
Announcements
Gad Allon, Northwestern University,
, Achal Bassamboo, Qiuping Yu
We study in a service environment, how to manage customers’ expectations and
to prioritize customers appropriately to maximize the firm’s profits. Specifically,
we focus on a setting where the firm uses only delay announcements and study
the opportunities and limitations of this mechanism. We are particularly
interested in when and how the customers can be influenced by delay
announcements.
2 - Trading Time In A Congested Environment
Luyi Yang, The University of Chicago, Chicago, IL, United States,
Laurens Debo, Varun Gupta
We propose time trading mechanisms, in which customers who are privately
informed about their waiting costs mutually agree on the ordering in the queue
by trading positions. We design optimal mechanisms for the social planner, the
service provider, and an intermediary who might mediate the trading platform.
Both the social planner’s and the service provider’s optimal mechanisms involve a
flat admission fee and an auction that implements strict priority. If a revenue-
maximizing intermediary operates the trading platform, it should charge a trade
participation fee and implement an auction with some restrictions on customer
trade.
3 - Learning To Bid In Sequential Auctions With Budgets Without
Market Information
Yonatan Gur, Stanford University,
Santiago Balseiro
We consider the of bidding in sequential auctions with budget constraints under
incomplete information, where bidders do not know the valuation distributions
and the budgets of their competitors. We present a general adaptive strategy
consisting of an approximation scheme in the dual space, in which bidders adjust
their multipliers (and accordingly, their bid functions) through the campaign
according to their expenditures. When adopted by all bidders, we show that the
long-run average payoffs of the strategy asymptotically converge to those under a
fluid mean-field equilibrium. We also analyze off-equilibrium performance under
arbitrary and utility maximizing deviations.
4 - Impact Of Uncertainty About Co-workers Capability On Server
Behavior In Queueing Systems
Masha Shunko, University of Washington, ,
Yaroslav Rosokha, Saurabh Bansal
We study business environment in which multiple servers process individual
orders, but receive payment based on the team performance. Specifically, we are
interested in the case when there is uncertainty about each other’s capability,
such as when teams are newly formed or when new members join the group.
What impact does this uncertainty have on the productivity of workers? Do
workers perform better or worse if they know what the co-workers are capable
of? Can the productivity be manipulated (impacted) by provision of relevant
information regarding others’ capability? We answer these questions using a
behavioral lab experiment.
MD43
208A-MCC
Simulation for Supply Chain
Sponsored: Simulation
Sponsored Session
Chair: Abdullah A Alabdulkarim, Dean, College of Engineering,
Majmaah University, university campus, Majmaah, Majmaah, 1176,
Saudi Arabia,
1 - Statistical Selection Of The Best Path In A Supply Chain
David Goldsman, Georgia Institute of Technology,
We study statistical formulations of problems involving the selection of the best
path from an origin to a destination through a network such as a supply chain.
The term “best” can take a variety of meanings, e.g., the path having the (i)
smallest expected travel time, (ii) highest probability of meeting a deadline, and
MD41