INFORMS Nashville – 2016
223
(iii) highest probability of yielding the shortest travel time. In each case, the goal
is to determine the number of experiments that one has to perform in order to
satisfy a certain indifference-zone requirement on the probability of correctly
selecting the best path. The experiments can involve actual observed travel times
or simulated realizations of travel times.
2 - Production Scheduling Of Jobs Limited By The Number Of
Machines And Workers With Setup Times And Route Flexibility
Alejandro Garcia del Valle, University of A Coruna,
Head Modeling & Simulation UMI Navantia, Ferrol, Spain,
alejandro.garcia.delvalle@udc.es, Javier Faulin,
Jose Antonio Muina Dono
Planning in dual resource constrained job shops like shipyards, is a very difficult
task, due to the large number of tasks and limited resources and the setup times.
In this paper, we develop a methodology using discrete event simulation models,
together with efficient and robust dispatching rules. This models are going to be
applied in the Navantia shipyard (Ferrol, Spain) in the context of a virtual
shipyard following the concept of Industry 4.0. Results will be presented
explaining the influence of jobs priority, assignation of workers to machines, and
the reduction of setup times.
3 - Stress Testing For Supply Chain Risk Management Using
Simulation Modeling
Harrison Luvai, University of Missouri - St. Louis, Saint Louis, MO,
63005, United States,
hl6d9@umsl.edu, L. Douglas Smith”
We employ a discrete-event simulation model with embedded MILP and
heuristics to test the resilience of a three-tier supply network when a blend of
strategies is used for mitigating risk. Production and material flows are adjusted in
response to daily simulated events. We investigate supply-network vulnerability
to system contagion by adjusting correlations in demands downstream and in
material deliveries upstream.
4 - MetaSimLab: A Laboratory For Validating And Calibrating
Agent-based Simulation For Business Analytics
Janina Knepper, Research Group Advanced Analytics,
Janina.knepper@ada.rwth-aachen.deAgent-based simulations are frequently used to develop and evaluate new and
improved approaches for business analytics and decision support. To be reliable,
they have to be empirically calibrated and validated. Existing calibration
approaches are rarely automated; however, manual calibration is costly in terms
of time and effort. Therefore, this contribution introduces the laboratory
environment MetaSimLab, which is designed to evaluate the efficiency and
effectiveness of alternative calibration approaches. We present numerical
examples to illustrate MetaSimLab’s functionality, novel calibration methods,
and an outlook on further research.
MD44
208B-MCC
DAS Awards Session
Sponsored: Decision Analysis
Sponsored Session
Chair: Eric Bickel, Associate Professor & Director, The University of
Texas at Austin, 204 E Dean Keeton St, Stop C2200, Austin, TX, 78712,
United States,
ebickel@utexas.edu1 - DAS Student Paper Award
Emanuele Borgonovo, Bocconi University, Milano, Italy,
emanuele.borgonovo@unibocconi.it,Robert Hammond
The Student Paper Award is given annually to the best decision analysis paper by
a student author, as judged by a panel of the Decision Analysis Society of
INFORMS. Students who did not complete their Ph.D. prior to May 1, 2015 are
eligible for this year’s competition.
2 - DAS Publication Award
Kenneth Charles Lichtendahl, University of Virginia,
lichtendahlc@darden.virginia.eduThis award is given annually to the best decision analysis article or book
published in the second preceding calendar year (i.e. calendar year 2014 for
consideration in 2016). The intent of the award is to recognize the best
publication in “decision analysis, broadly defined.” This includes, but is not
limited to, theoretical work on decision analysis methodology (including
behavioral decision making and non-expected utility theory), descriptions of
applications, and experimental studies.
3 - DAS Practice Award
Franklyn Koch, Koch Decision Consulting,
kochfg@gmail.comThe Decision Analysis Practice Award is awarded to the best example of decision
analysis practice as judged by the Decision Analysis Practice Award Committee.
The purpose of the award is to publicize and encourage outstanding applications
of decision analysis practice. We will present the finalists and this year’s winner.
4 - DAS Ramsey Medal
Jeffrey M Keisler, University of Massachusetts - Boston,
jeff.keisler@umb.eduThe Ramsey Medal of the Decision Analysis Society is awarded for distinguished
contributions in decision analysis. Distinguished contributions can be internal,
such as theoretical and procedural advances in decision analysis, or external, such
as developing or spreading decision analysis in new fields. We will introduce the
2016 Ramsey Medal winner, followed by a presentation by the winner.
MD45
209A-MCC
Simulation-based Optimization
Sponsored: Simulation
Sponsored Session
Chair: Tahir Ekin, Texas State University, San Marcos, TX,
United States,
tahirekin@gmail.com1 - Nested Augmented Simulation For Stochastic Optimization
Tahir Ekin, Texas State University,
t_e18@txstate.edu, Refik Soyer,
Nick Polson
This talk presents nested augmented probability simulation to solve for decision
making problems with uncertainty. The focus will be on stochastic programs with
recourse under decision dependent (endogenous) uncertainty. Augmented
probability simulation is based on the idea of treating the decision variable as
random and investigating the optimal decision in the joint space of decision and
random variables. We present the use of Nested Sampling for simulation from this
joint distribution. An illustration is provided on a two stage news-vendor
problem. We provide performance comparisons with traditional Monte Carlo
simulation and present computational insights.
2 - Bayesian Inference And Augmented Probability Simulation In
Call Center Staffing
Tevfik Aktekin, University of New Hampshire,
Tevfik.Aktekin@unh.eduWe consider the issue of short term staffing in a queuing system such as a call
center where the system rates are dependent random variables. We consider their
estimation using Bayesian inference and the well-known Erlang A queueing
model. We formulate the optimization model such that both the objective
function and the constraints are random due to the uncertainty in the system
rates and propose the use of an augmented probability simulation approach. In
our numerical illustration, we consider both real and simulated data examples. In
each case, we divide the day into discrete time intervals to determine staffing
levels and discuss further implications of our approach.
3 - A Simulation Optimization Framework For Scheduling Preventive
Maintenance In Wind Energy Systems
Eduardo Perez, Texas State University,
eduardopr@txstate.eduBecause of the continuously escalating costs of wind farms O&M in the United
States, determining methods for using available data in conditioning-monitoring
systems is critical to decreasing wind farms operational costs. To accomplish this
objective, we have developed a data-driven integrated stochastic online
optimization and discrete event simulation methodology that takes into account
data uncertainties in turbines status, weather conditions, and resources
availability in scheduling maintenance and resources. Discrete event simulation
coupled with optimization provides a powerful instrument for assessing and
revising schedules prior to actual implementation.
4 - On The Sample Average Approximation Of The Two Stage
Chance Constrained Staffing Problem In Call Centers With Arrival
Rate Uncertainty
Anh Thuy Ta, PhD Student, University of Montreal, CP 6128
Succursale Centre-Ville, Montreal, QC, H3W1C5, Canada,
tathuyanh1989@gmail.com, Wyean Chan, Pierre L’ecuyer,
Fabian Bastin
We consider a chance constrained two stage stochastic staffing problem for multi-
skill call centers with arrival rate uncertainty. The aim is to minimize the total cost
of agents under some chance constraints, defined over the randomness of the
service level in a given time period. We use the Monte Carlo method to generate
M scenarios of arrival rates and we perform N simulation runs to get the estimates
of probabilities that the service level is satisfied. We then obtain a sample average
approximation (SAA) of the problem. We investigate the convergence of the
optimal solution of the SAA to that of the original one when the sample size
increases and present numerical illustrations on the sample sizes M and N.
MD45