INFORMS Nashville – 2016
180
Using Anom Slicing For Multiway Models With Binomial
Or Poisson Data
peter wludyka, CEO, Wludykaandassociates, 4285 Baltic Street,
Jacksonville, FL, 32210, United States,
pwludyka@unf.edu,
John Noguera
Results from “Using ANOM Slicing for Multi-Way Models with Significant
Interaction” (Wludyka, JQT 2015) are extended to binomial and Poisson data.
Analysis of Means (ANOM) slice charts using an overall estimate for the
proportion/rate (which will often have more power than by analysis charts) are
presented.
Data-driven Decision Making In The Sharing Economy
Qiaochu He, Assistant Professor, University of North Carolina,
Charlotte, 9201 University City Blvd., Univ. North Carolina,
Charlotte, NC, 28223, United States,
qhe4@uncc.eduYun Yang
In this presentation, we propose several new models related to the sharing
economy. We investigate this industry from a data-driven perspective, and focus
on the following issues: (1) Delayed matching mechanism in a two-sided market;
(2) The value of forecasting in matching with uncertainty; (3) Emerging service
mechanism in the sharing economy.
Optimization Problems In The Design And Control Of Internet
Fulfillment Warehouses
Sevilay Onal, PhD Candidate, New Jersey Institute of Technology,
323 Dr Martin Luther King Jr Blvd, Newark, NJ, 07102,
United States,
so59@njit.edu,Jingran Zhang, Sanchoy Das
Internet Fulfillment Warehouses (IFWs) are facilities that have been designed and
built exclusively to process online retail orders. The nature of e-commerce is
described with a high number of transactions in small quantities. Therefore,
human controlled systems are being replaced with total digital control. Thus, a
revision is required for existing methodology such as adoption of pick and storage
policies, resource allocation strategies, warehouse design and control. We aim to
introduce the operational and design environment of IFWs and summarize
associated emerging optimization problems
Container Port Selection In West Africa A Multicriteria
Decision Analysis
Rivelino De Icaza, PhD Candidate, University of Arkansas,
1301 N Prairie Dunes Trail, Apt 305, Fayetteville, AR, 72704,
United States,
rdeicaza@uark.eduWest Africa gross domestic product is expected to grow to 6.2 percent in 2016 and
port capacity will increase by over 12 million TEUs by 2020. Despite the region
economic potential and the steady grow of container traffic over the years, port
selection decision by shipping lines is complex because the region still has poor
shipping infrastructure and political instability that impact transportation security,
and consequently the logistics and supply chain services. This research applies a
multiattribute value theory (MAVT) with valued-focused thinking (VFT) and an
alternative-focused thinking (AFT) methodologies to develop a shipping lines’
container port selection decision models.
Dynamic Data-driven Physician Rostering Under
Variable Availability
Monique Bakker, City University of Hong Kong, Hong Kong SAR,
Hong Kong.
mbakker2-c@my.cityu.edu.hk, Kwok L. Tsui
Efficient staff rostering and patient scheduling to meet outpatient demand is a
very complex and dynamic task. Medical specialists are typically restricted in sub-
specialization, serve several patient groups and are the key resource in a chain of
patient appointments at the outpatient clinic, endoscopy unit, and surgical unit.
We present a new, data-driven algorithmic approach to automatic allocation of
specialists to activities and patient groups. This approach minimizes variability in
specialist activity rosters. It outperforms traditional cyclic scheduling with
increased patient service level (% patients served in time) and capacity utilization,
and decreased patient wait time (days).
How Can Mathematical Modelling Quantify Future Fishing Risks
Under Climate Change Scenarios?
Sara Rezaee, Dalhousie University, Halifax, NS, Canada, BC,
Canada.
sara.rezaee@dal.ca, Ronald P. Pelot, Christian Seiler,
Alireza Ghasemi
Studies have shown that extreme weather factors can affect fishing safety
significantly. Changes in weather patterns due to climate change effects will add
uncertainty to fishing safety systems. This study proposes a framework to quantify
fishing incident risks in the future due to changes in weather conditions. The
framework builds relationships between fishing incidents and weather conditions
based on historical data using mathematical modelling and data mining
techniques and then predicts future risks according to these relationships with
respect to potential changes in weather patterns.
Evaluation Of Low Carbon Level Of Enterprise Logistics Based
On Improved Entropy Law And Cloud Model
Mi Gan, Southwest Jiaotong University, Chengdu, China.
migan@swjtu.cn,Xiaofan Guo, Shuai Yang, Lei Wang
To evaluate the low-carbon level of logistics enterprise, we construct
corresponding evaluation index system and introduce the concept of centrality in
DEMATEL method to improve the entropy method, use data variability to do
objective weighting of evaluation index. Then combine with cloud model theory,
use reverse cloud generator to convert evaluation data into cloud parameters, and
use index approximation method to construct evaluation cloud, reflect and obtain
the evaluation results in the form of normal cloud chart, to solve fuzziness and
randomness quantification in the evaluation process.
Poster Competition
Exhibit Hall
Monday Poster Competitition
Competition Poster Session
A Simple Classification Framework For Discrimination Of
Antipsychotic Treatment Resistant And Treatment Responsive
Schizophrenia Patients
Farnaz Zamani Esfahlani, PHD Candidate, SUNY Binghamton,
4400 Vestal Parkway East, Binghamton, NY, 13902-6000,
United States,
fzamani1@binghamton.edu, Katherine Frost,
Gregory Strauss, Hiroki Sayama
Predicting the outcome of different treatment options for patients is essential for
effective treatment planning. However, this is a challenging task especially in
mental disorders such as schizophrenia where the treatment outcome of patients
depends on the complex interaction of various symptoms. In this study, we used
analytical tools of network science to study the interaction of the symptoms in
schizophrenia and identify symptoms that best discriminate antipsychotic
treatment resistant and treatment responsive schizophrenia patients. The features
selected based on network analysis provided better classification accuracy when
compared to traditional feature selection methods.
Anticipatory Dynamic Traffic Sensor Location Problems with
Connected Vehicle Technologies
John (Hyoshin) Park, Research Associate, University of
Massachusetts Amherst, 370 Northampton Rd. Apt B, Amherst,
MA, 01002, United States,
hyoshin0724@gmail.com,Song Gao,
Ali Haghani
Despite the potential benefits of sensor technologies, the challenges associated
with identifying optimal sensor locations for multiple time stages throughout a
day with uncertain demand patterns has received little attention. In this paper,
we focus on proactively reducing the network delay by controlling traffic signals
through an optimized sensor deployment. The framework is based upon portable
sensors that may be repositioned within the day to new locations such that delay
savings over multiple time stages will be maximized. To tackle this multi-period
stochastic problem, dynamic models are proposed, considering the future sensor
locations given budget constraints on the sensor costs and relocation costs. A
subproblem decomposed by Lagrangian relaxation enhanced with valid cuts has a
better bound and a variable neighborhood search algorithm quickly finds
solutions. Two dynamic models that constrain a flexible or restricted relocation
present higher savings compared to the stationary model without sensor
relocation. The flexible relocation model guarantees higher savings than restricted
model by achieving the same maximum savings with fewer number of sensors.
Resilient Offgrid Microgrids – Capacity Planning And N-1 Security
Sreenath Chalil Madathil, Clemson University, 324 Village Walk
Ln, Clemson, SC, 29631, United States,
schalil@g.clemson.eduScott J. Mason, Russell Bent, Harsha Nagarajan, Emre Yamangil,
Scott Backhaus, Arthur Barnes, Salman Mashayekh,
Michael Stadler
Despite the long distance power transmission capabilities, there are some remote
communities in Alaska and Hawaii that are not connected to these systems. These
communities rely on small, disconnected microgrids to deliver power. These
microgrids are not held to same reliability standards as transmission grids and can
place many communities at risk for extended black-outs. To address this issue, we
develop an optimization model and algorithm for capacity planning and
operations of microgrids that includes N-1 security and other modeling features.
The effectiveness of the approach is demonstrated using the IEEE 13 node test
feeder and a model of the Nome, Alaska distribution system.
POSTER COMPETITION