INFORMS Nashville – 2016
423
4 - Biased Randomization: Heuristics In Transportation, Logistics,
And Production
Alex Grasas, EADA Business School, C/ Arago 204, Barcelona,
08011, Spain,
agrasas@eada.edu, Angel A Juan, Javier Faulin,
Jesica De Armas, Helena Ramalhinho
This paper reviews heuristics that contain biased-randomized procedures (BRPs).
A BRP is a procedure that uses a biased probability distribution to select the next
constructive movement at each algorithm’s iteration. BRPs can be categorized into
two main groups according to how choice probabilities are computed: (i) BRPs
using an empirical bias function; and (ii) BRPs using a skewed probability
distribution. This paper analyzes the second group and reviews the use of these
BRPs in some applications in transportation, logistics, and production problems.
5 - A Parallel Dynamic Programming Solution For The Dynamic
Facility Layout Problem
Clara Novoa, Associate Professor, Texas State University,
601 University Dr, San Marcos, TX, 78666, United States,
cn17@txstate.edu, Apan Qasem, Chandra Kolla
We develop a parallel approximate dynamic programming solution to the
Dynamic Facility Layout Problem (DFLP) using OpenMP. We experiment with
data sets from Dr. Balakrishnan’s repository. Including a relatively small set of
feasible solutions, the accuracy and speed of our method is very satisfactory if
contrasted to simulated annealing, hybrid genetic algorithm, and ant systems. In
the DFLP, the flow of materials between departments is known but it varies over
time due to changes in demand and introduction of new products. The trade-off
costs are material handling and relocation costs.
WB79
Legends G- Omni
Opt, Convex
Contributed Session
Chair: Churlzu Lim, Associate Professor, University of North Carolina at
Charlotte, Systems Engineering & Engineering Management, 9201
University City Boulevard, Charlotte, NC, 28223, United States,
clim2@uncc.edu1 - A Fast Socp-based Method For Optimal Selection Problem In
Tree Breeding
Makoto Yamashita, Associate Professor, Tokyo Institute of
Technology, W8-29 2-12-1 Oh-Okayama, Meguro, Tokyo,
152-8552, Japan,
makoto.yamashita@c.titech.ac.jp, Tim J Mullin,
Sena Safarina
One of new frontiers for optimization methods is to solve practical problems
arising from breeding. A purpose of an optimal selection problem in tree breeding
is to determine the contributions of candidate genotypes that attains the highest
profit subject to a constraint on genetic diversity. We propose a fast numerical
method based on second-order cone programming by exploiting the structural
sparsity in the problem. The proposed method reduced the computation time
from 39,000 seconds of an existing method to just 2 seconds.
2 - Shape Constrained Data Smoothing With Penalized Splines
Yu Xia, Professor, Lakehead University, Business Administration,
955 Oliver Rd, Thunder Bay, ON, P7B 5E1, Canada,
yxia@lakeheadu.ca, Farid Alizadeh
We consider fitting noisy data to a smooth function by penalized B-splines. The
underling function is assumed to have some shape properties, such as non-
negative, monotonic, convex. We solve the data smoothing problem by convex
optimization methods.
3 - On The Convexity Of Optimal Decentralized Control Problem And
Sparsity Path
Salar Fattahi, PhD Student, University of California, Berkeley,
4141 Etcheverry Hall, Berkeley, CA 94720-1777, Berkeley, CA,
94702, United States,
fattahi@berkeley.edu, Javad Lavaei
This talk is about an important special case of the optimal stochastic decentralized
control problem, where the objective is to design a static structured controller for
a stable stochastic system. We show that if either the noise covariance or the
input weighting matrix is not too small, the problem is locally convex. In the case
where these conditions are not satisfied, we modify the problem by a penalization
term to convexify it, leading to a near-global solution. We also study the problem
of designing a sparse controller using a regularization technique. Under some
genericity assumptions, we prove that this method is able to design a controller
with any arbitrary sparsity level.
4 - Snug Circumscribing Simplexes For Convex Hulls
Ghasemali Salmani Jajaei, PhD Student, Virginia Commonwealth
University, 1015 Floyd Avenue, Harris Hall, Richmond, VA,
23284-3083, United States,
salmanijajaeg@vcu.eduWe propose procedures for enclosing convex hulls of finite m-dimensional point
sets with simplexes. These are snug in since they intersect the hull in some way.
We report on experimental results.
5 - Volume Allocation Optimization For Space Mission Tasks
Churlzu Lim, Associate Professor, University of North Carolina at
Charlotte, Systems Engineering & Engineering Management, 9201
University City Boulevard, Charlotte, NC, 28223, United States,
clim2@uncc.edu,Simon M Hsiang, Sherry Thaxton, Maijinn Chen,
Jerry G Myers
Volume in space missions is costly and often must be traded with competing
resources and mission needs, such as launch mass, systems/hardware
requirements, and consumables. Spacecraft and habitat designers must allocate
sufficient volume for different tasks without incurring excessive cost penalties or
failure modes. How a volume can be optimized should be based on a balancing of
risk and benefits. In this talk, we present a mathematical optimization model that
maximizes the total value of tasks of astronauts given a limited volume in the
spacecraft. An illustrative example will be demonstrated.
WB80
Broadway E- Omni
Health Care, Public II
Contributed Session
Chair: Samantha Meyer, Assistant Professor, Ross School of Business,
701 Tappan Ave., Ann Arbor, MI, 48109, United States,
srmeyer@umich.edu1 - How Much Sleep Do You Need?: Evidence From Public Health
Philip F. Musa, Associate Professor and Programs Director,
University of Alabama-Birmingham, PO Box 55544, Birmingham,
AL, 35255, United States,
musa@uab.eduCould the amount of sleep people get be associated with hypertension? This
presentation outlines an epidemiological cross-sectional study to shed some light
on this important Public Health chronic matter. We present a background from
the literature using a population based sampling. Our proposed study will employ
the previously validated Pittsburgh Sleep Quality Index and Berlin questionnaire.
The inclusion/exclusion criteria and the strengths and limitations are presented.
2 - A Little Empathy Goes A Long Way In Disease Dynamics On A
Network Game
Ceyhun Eksin, Georgia Tech, 85 5th St. NW, Atlanta, GA, 30332,
United States,
ceyhuneksin@gatech.edu,Jeff S Shamma,
Joshua Weitz
Individuals change their behavior during an epidemic in response to whether
they and those they interact with are healthy or sick. Healthy individuals may
utilize protective measures while sick individuals may adopt preemptive measures
to stop disease spread to their contacts. Yet, in practice both protective and
preemptive behavior come with costs. We propose a stochastic network disease
game that captures the self-interests of individuals during disease spread on a
network. We show that there is a critical level of concern, i.e., empathy, by the
sick individuals that eradicates the disease fast while the protective measures
cannot eradicate the disease without the preemptive measures.
3 - Spatial Evolutionary Game For Changes Of Human Behaviors
In Epidemic
Songnian Zhao, Kansas State University, 2037 Durland Hall,
Manhattan, KS, 66502, United States,
songnian@ksu.edu,Yan Kuang, John C Wu, David Ben-Arieh
Spatial evolutionary game, used to study multiple players’ behaviors in a spatial
structure, was incorporated into the epidemic models for sake of evaluating
spontaneous changes of human behaviors when individuals acquire information
about the spread of infectious disease and make a tradeoff between costs and
benefits. Through the comparison between two different mechanisms, a spatial
game model in epidemic is validated to generate the consistent results with the
traditional dynamic systems in this paper.
WB80