2016 INFORMS Annual Meeting Program

134 / 561

INFORMS Nashville – 2016
134
MA37
205C-MCC
Securing Sustainable Future
Sponsored: Manufacturing & Service Oper Mgmt, Sustainable
Operations
Sponsored Session
Chair: Elena Belavina, University of Chicago Booth School of Business,
Chicago, IL, United States,
1 - Grocery Access Market Structure And Food Waste
Elena Belavina, University of Chicago,
Access to grocery, or how dense is the network of retail stores in a neighborhood,
varies extensively as a result of zoning laws and other city government initiatives.
Similarly, some markets are dominated by one chain, while others have a high
degree of competition with a lot of independent grocery stores. This paper studies
how access to grocery stores, and the extent and nature of competition in the
grocery retail market influences food waste.
2 - Pricing, Product Display, Inventory And Waste Management For
Deteriorating Products.
Dorothee Honhon, University of Texas at Dallas,
Richardson, TX, United States,
Xiajun Amy Pan, Zumbul Atan
We consider the problem of a retailer managing the inventory and the prices of
products whose quality deteriorates over time. We show that by appropriately
displaying the most/least fresh products on the store shelves, the retailer can, in
some cases, increase profits and reduce waste.
3 - The Impact Of Legislation on Food Waste
Alexandra Heeney, Stanford University, Stanford, CA, 9,
United States, Warren H Hausman,
Erica Plambeck
Worldwide, 30-40% of all food is wasted, which has a significant environmental
impact: agriculture is responsible for 22% of all greenhouse gas emissions. This
research explores whether California’s AB32 and renewable energy legislation
(that exempts farmers from carbon tax) increases the negative environmental
impact of the food system (by preventing the true cost of food to propagate down
the supply chain making food waste relatively inexpensive). Further, we study
how the supply chain structure and design contribute to food waste and strategies
for mitigating this.
4 - Design Implications Of Extended Producer Responsibility For
Durable Products
Ximin (Natalie) Huang, Georgia Institute of Technology, 800 West
Peachtree Street, NW, Atlanta, GA, 30308, United States,
, Atalay Atasu, Beril L Toktay
We consider a monopolist who has two product design options to manage the
end-of-life costs/revenues associated with its products: making products more
durable or recyclable. We explore how the recyclability and durability choices are
affected by the requirements of take-back legislation.
MA38
206A-MCC
Meet the Editors Panel – NPD, Innovation,
and Technology
Invited: New Product Development
Invited Session
Moderator: Sanjiv Erat, University of California-San Diego, UCSD,
La Jolla, CA, United States,
1 - Meet The Editors Panel - NPD, Innovation, andTechnology
Joel Wooten, University of South Carolina, 1014 Greene St.,
Columbia, SC, 29208, United States,
This interactive session aims at assisting readers and researchers in staying
informed on the most important topics and the latest development in New
Product Development, Technology, and Innovation Management.
2 - Panelist
Cheryl Gaimon, Georgia Institute Of Technology, Scheller College Of
Business, Atlanta, GA, 30308, United States,
3 - Panelist
Moren Levesque, York University, Schulich School Of Business,
Toronto, ON, Canada,
MA39
207A-MCC
Applied Probability and Machine Learning II
Sponsored: Applied Probability
Sponsored Session
Chair: Sewoong Oh, UIUC, 2011 Savanna Dr, Champaign, IL, 61822,
United States,
1 - Online Rules For Control Of False Discovery Rate
Adel Javanmard, Assistant Professor, University of Southern
California, Bridge Memorial Hall, 3670 Trousdale Parkway, Los
Angeles, CA, 90089, United States,
Andrea Montanari
Multiple hypothesis testing is a core problem in statistical inference and arises in
almost every scientific field. A common error criteria in this context is the false
discovery rate (FDR). In this talk, we consider the problem of controlling FDR in
an “online manner”. Concretely, we consider an ordered, possibly infinite,
sequence of null hypotheses where at each step the statistician must decide
whether to reject current null hypothesis having access only to the previous
decisions. We introduce a class of generalized alpha-investing procedures and
prove that any rule in this class controls FDR in online manner. Time permitting,
we will discuss applications for Ad click predictions and A/B testing.
2 - Newton Stein Method: An Optimization Method For Glms
Murat Erdogdu, Stanford University,
We consider the problem of efficiently computing the maximum likelihood
estimator in Generalized Linear Models (GLMs) when the number of observations
is much larger than the number of coefficients (n>>p>>1). In this regime,
optimization algorithms can immensely benefit from approximate second order
information. We propose an alternative way of constructing the curvature
information by formulating it as an estimation problem and applying a Stein-type
lemma, which allows further improvements through sub-sampling and
eigenvalue thresholding. Our algorithm enjoys fast convergence rates, resembling
that of second order methods, with modest per-iteration cost.
3 - Data-driven Rank Breaking For Efficient Rank Aggregation
Ashish Khetan, UIUC, Urbana, IL, United States,
Rank aggregation systems collect ordinal preferences from individuals to produce
a global ranking. Rank-breaking is a common practice to reduce the
computational complexity of learning the global ranking. The individual
preferences are broken into pairwise comparisons and applied to efficient
algorithm. However, naive rank-breaking approaches can result in inconsistent
estimates. The key idea to produce accurate and unbiased estimates is to treat the
pairwise comparisons unequally, depending on the topology of the collected data.
In this paper, we provide the optimal consistent rank-breaking estimator. This
allows us to characterize the trade-off between accuracy and complexity.
4 - On The Capacity Of Information Processing Systems
Kuang Xu, Stanford University, Stanford, CA, United States,
Laurent Massoulie
We analyze a family of information processing systems, where a finite set of
experts or servers are employed to extract information about a stream of
incoming jobs, each associated with a hidden label. An inspection by an expert
produces a noisy outcome that depends both on the job’s hidden label and the
type of the expert, and occupies the expert for a finite time duration. A decision
maker’s task is to dynamically assign inspections so as to accurately recover all job
labels while keeping the system stable. Crowdsourcing, diagnostics and
experiment designs are among our chief motivations. Our main result is an
asymptotically optimal inspection policy that utilizes the fewest experts.
MA37