2016 INFORMS Annual Meeting Program

35 / 561

INFORMS Nashville – 2016
35
3 - Information Dissemination Through Social Media In
Humanitarian Operations
Eunae Yoo, Arizona State University, Tempe, AZ, United States,
, Elliot Rabinovich, Bin Gu, William Rand,
Mahyar Eftekhar
For humanitarian operations, the distribution of information is critical to support
the effective and efficient delivery of goods and services. Since social media
facilitates real-time information sharing, humanitarian organizations have started
leveraging these platforms to communicate with their stakeholders. Our research
examines information diffusion patterns on social media during disasters, taking
into account the underlying social network among users. From our results, we
identify how humanitarian organizations can improve information propagation
on social media in emergencies.
4 - Disaster Cycle Management: Matching Supply And Demand Of
Social Support Through Social Media.
Alfonso Pedraza-Martinez, Indiana University,
Lucy Yan
We study information management during the disaster cycle. This research
investigates the match of social support supply provided by organizations and
social support demand from social media users during the stages of preparedness,
response and early recovery. Using the case of Hurricane Sandy, 2012 we find a
mismatch between supply and demand for social support and provide discussions
on alleviating the mismatch. Moreover, we study how actionable operations
management content posted by organizations affects their interaction with users
through social media.
SA52
214-MCC
Prevailing Issues in Public Sector OR
Sponsored: Public Sector OR
Sponsored Session
Chair: Ebru Bish, Virginia Tech, Dept of Industrial and Systems
Engineering, Blacksburg, VA, 24061, United States,
1 - The Probabilistic Independence Of Mass Killings In The
United States
Douglas M King, Lecturer, University of Illinois at Urbana-
Champaign, 117 Transportation Building, 104 S Mathews Avenue,
Urbana, IL, 61801, United States, ,
Sheldon Jacobson
As mass killings are increasingly visible in media coverage, understanding
patterns in their occurrence becomes increasingly important. This talk finds that
the 300 mass killings taking place in the United States from 2006 to 2016, as
documented by USA Today, have a temporal distribution indistinguishable from a
homogeneous Poisson process. This result suggests that these events are
independent in that the occurrence of one mass killing does not indicate whether
another is imminent.
2 - Adaptive Array-based Screening For Heterogeneous Populations
Hrayer Aprahamian, Virginia Tech, 250 Durham, Blacksburg, VA,
24061, United States, , Ebru Korular Bish,
Douglas R Bish
Group (pooled) testing has seen many applications, especially in the context of
public health, blood screening, and genetics. We consider a special form of group
testing called “array-based testing,” which takes advantage of overlapping pools.
We model an adaptive and informative testing scheme that considers important
test and population characteristics, including imperfect tests, dilution effect, and
heterogeneity in the population, and determine the structure of the optimal
testing design and optimal assignment of the heterogeneous subjects to the pools.
Our case study indicates that the proposed optimization-based model leads to a
substantial improvement over current practices.
3 - Prevalence Estimation Through Pooled Testing
Ngoc Nguyen, Virginia Tech, Ebru Korular Bish,
Douglas R Bish
Surveillance studies often rely on pooled testing in order to estimate the
unknown prevalence rate of an infection or a genetic disorder. Utilizing larger
pools reduces testing costs, but also leads to a loss of accuracy due to dilution
effects. We develop a mathematical model that considers this trade-off and
determines the optimal pool size as well as the optimal number of pools under
resource constraints. Our case study shows that this optimization-based approach
improves upon the accuracy of the prevalence rate estimate over current
approaches.
4 - Production And Distribution Capacity Planning For Mitigating
Urban Delivery Risk Of E-commerce
Mu Du, Dalian University of Technology, 509 School of
Management, No.2 Linggong Road, Dalian, 116023, China,
, Nan Kong
The integrated production and distribution capacity planning is vital to the supply
chains of perishable commodity B2C e-commerce, which are keen on reducing
production makespan and delivery lateness. However, significant challenge lies in
the uncertain delivery workload caused by traffic control and adverse weather
(e.g., smog) in Chinese urban areas. We formulate a two-stage stochastic
programming model and propose a stochastic branch and bound algorithm. We
report the impact of traffic control and adverse weather on the capacity planning
decisions, confounded by other factors.
SA53
Music Row 1- Omni
Topics in Revenue Generation from Innovation
Sponsored: Technology, Innovation Management &
Entrepreneurship
Sponsored Session
Chair: Pascale Crama, Singapore Management University, 50 Stamford
Road, Singapore, 178899, Singapore,
1 - Retaining Capable New Employees: Role Of Strategic Interaction
And The Learning Rate
Onesun Steve Yoo, University College London, London,
United Kingdom, , Dharma Kwon
We study a two-sided game involving a firm and a newly hired employee whose
capability is not initially known to either party. As the employee performs, both
players learn and are presented with an option: the firm can terminate an
incapable employee, and a capable employee can leave the firm for greater
financial remuneration elsewhere. We examine the Markov perfect equilibrium
(MPE) termination strategies and payoffs that unfold. We report a
counterintuitive result: slower learning can increase the equilibrium payoff for
both parties. Our result identifies a nonfinancial way for firms to improve
retention of highly capable employees and create a win-win situation for both
parties.
2 - The Impact Of Valuation Heterogeneity And Network Structure On
Equilibrium Prices In Supply Networks
Alper Nakkas, Nova School of Business and Economics, Campus de
Campolide, VAT - 506030636, Lisbon, 1099-032, Portugal,
, Yi Xu
We study how valuation heterogeneity and network structure on equilibrium
prices in supply networks by identifying the main factors that influence the
equilibrium prices, trading pattern and surplus allocation in such networks. We
also show what types of links can be added into a supply network to improve its
competitiveness and/or efficiency.
3 - Signaling Product Quality Through A Trial Period
Gulru Ozkan-Seely, University of Washington – Bothell, Bothell,
WA, 98033, United States, , Shouqiang Wang
We evaluate a firm’s pricing and timing strategies when launching via a time-
locked trial period a new product that has privately observable product quality.
While the trial period allows consumers to learn about the product quality (a
phenomenon we term a learning effect), a longer trial period increases the
heterogeneity of consumers’ willingness-to-pay for the firm (a phenomenon we
term a dispersion effect). The dispersion effect exacerbates the firm’s pricing
difficulty and generates an informational cost, warranting a longer trial period as
a credible signal of the firm’s superior product quality. In a key finding, we show
that a firm can use the price and the trial length as dual signals.
4 - An Experimental Study Of Idea Evaluation Process
Zhijian Cui, IE Business School,
With an online experiment, this study compares the efficacy of two idea
evaluation processes commonly observed in practice: ranking and scoring. We
find that the scoring process has a higher evaluation accuracy than the ranking
process. In addition, providing more information and domain-specific expertise
could improve the evaluation accuracy of scoring process, but not the ranking
process.
SA53