2015 Informs Annual Meeting

MD34

INFORMS Philadelphia – 2015

3 - A Chance Constrained Programming Approach to Handle Uncertainties in Radiation Treatment Planning Maryam Zaghian, University of Houston, 4800 Calhoun Rd, Houston, TX, United States of America, mzaghian@uh.edu, Azin Khabazian, Gino Lim A chance constrained programming (CCP) framework is developed to handle set- up uncertainties in radiation treatment planning. By allowing some degree of violations of constraints, the proposed approach optimizes the treatment plan while satisfying the planner’s tolerance level on the constraint violation in a probabilistic environment. Linear deterministic equivalences of the chance constraints are derived under distributional assumptions on uncertainties. MD34 34-Room 411, Marriott Decision Models for Women’s and Children’s Health Sponsor: Health Applications Sponsored Session Chair: Karen Hicklin, PhD Student, North Carolina State University, 111 Lampe Drive, Campus Box 7906, Raleigh, NC, 27695, United States of America, kthickli@ncsu.edu 1 - Modeling Comorbidity in Women with Diabetes Nisha Nataraj, PhD Student, North Carolina State University, 111 Lampe Drive, Campus Box 7906, Raleigh, NC, 27695, United States of America, nnatara@ncsu.edu, Fay Cobb Payton, Julie Ivy Comorbidity is the presence of two or more concurrently existing conditions in an individual. A 2012 CDC report estimates that one in four US adults have comorbid conditions, contributing heavily to healthcare spending. Our focus is on diabetes since it is associated with significant comorbidity. Using National Inpatient Sample data (2006-2011), we build a modeling framework that helps evaluate how comorbidity impacts prognosis and outcomes for women with diabetes at a population level. 2 - Using Simulation to Determine a Balance between Cost and Quality of Care for Critically Ill Infants Emily Lada, Principal Operations Research Specialist, SAS Institute Inc., SAS Campus Drive, Cary, NC, United States of America, Emily.Lada@sas.com, Chris Derienzo, David Tanaka, Phillip Meanor Discrete-event simulation techniques are used to assess the relationship between cost, average length of stay, and patient outcomes in a neonatal intensive care unit (NICU). The model represents a general method that can be applied to any NICU, thereby providing clinicians and administrators with a tool to quantitatively support staffing decisions. Over time, the use of the model can lead to significant benefits in both patient safety and operational efficiency. 3 - A Bayesian Markov Decision Process to Evaluate Mode of Delivery for Laboring Women Karen Hicklin, PhD Student, North Carolina State University, 111 Lampe Drive, Campus Box 7906, Raleigh, NC, 27695, United States of America, kthickli@ncsu.edu, Fay Cobb Payton, Vidyadhar Kulkarni, Meera Viswanathan, Evan Myers, Julie Ivy A laboring woman will deliver through one of two ways: successful trial of labor or C-section. We combine Bayesian updating into a Markov decision process to determine under what circumstances it is appropriate to gather more information before making a decision regarding mode of delivery. The goal is to maximize the utility of health outcomes for the mother and child as a function of the belief that the woman will have a safe vaginal delivery as a function of cervical dilation progression. MD35 35-Room 412, Marriott Joint Session PPSN/Analytics: Pro Bono Analytics Panel Discussion Sponsor: Public Sector OR Sponsored Session Chair: David Hunt, Manager, Oliver Wyman, One University Square, Suite 100, Princeton, NJ, 08540, United States of America, David.Hunt@oliverwyman.com 1 - Pro Bono Analytics Panel Discussion Moderator:David Hunt, Manager, Oliver Wyman, One University Square, Suite 100, Princeton, NJ, 08540, United States of America, David.Hunt@oliverwyman.com, Panelists: Evan Fieldston, Joel Zarrow

Pro Bono Analytics (PBA) is a new initiative within INFORMS to match members willing to volunteer their skills with non-profit organizations working in underserved and developing communities. To launch PBA, representatives from prominent Philadelphia area non-profit organizations will participate in a panel discussion exploring the types of problems they face and ways that analytics/OR methods can help. Please join us to learn about the types of analytical problems at non-profits, and about PBA. MD36 36-Room 413, Marriott Modeling Broader Policy Impacts at the Local Scale Sponsor: Public Sector OR Sponsored Session Chair: Ronald McGarvey, Indust. & Manuf. Systems Engineering; Truman School Of Public Affairs, University of Missouri, 225 Maryam Alsadat Andalib, PhD Student, Virginia Tech, 536F Whittemore Hall, 1185 Perry Street, Blacksburg, VA, 24061, United States of America, maryam7@vt.edu, Vida Abedi, Arash Baghaei Lakeh, Ramin Zand, Navid Ghaffarzadegan, Niyousha Hosseinichimeh, Grant Hughes The accuracy of survey data for analyzing health policy problems depends on people’s willingness to admit and honestly respond to survey questions. In the mental health context, stigma is a barrier which affects accuracy of survey data. We investigate fidelity of using another data source, Google search queries, to understand mental health illnesses. We specifically offer three examples of analyzing mental health illnesses in the United States. 2 - Robust Optimization for Biopower Generation Bayram Dundar, University of Missouri-Columbia, 200 Engineering Building North, Columbia, MO, 65211, United States of America, bd5zc@mail.missouri.edu, Ronald McGarvey, Francisco X. Aguilar The U.S. Environmental Protection Agency (EPA) has proposed a rule that aims to reduce carbon emissions from coal-fired power plants. We develop an MILP model to identify min-cost approaches for satisfying these proposed standards via biopower generation subject to spatially-explicit biomass constraints. We next propose a robust optimization model to address parameter uncertainty, and compare the two models’ results to illustrate the impact of data uncertainty on overall cost and emissions. 3 - Modeling the Recruitment and Retainment of Employees at a Rural Montana Community Health Center Andreas Thorsen, Assistant Professor Of Management, Montana State University, 330 Jabs Hall, Bozeman, MT, 59717, United States of America, holger3000@gmail.com, Don Greer, Laura Black, Edward Gamble Community Health Centers (CHC) are not-for-profit health care corporations which provide comprehensive medical and dental care for their communities regardless of the individual’s insurance coverage or ability to pay. For a CHC in rural Montana, there are unique challenges related to recruitment and retention of highly qualified, mission-driven employees. We identify and address these challenges using a system dynamics modeling approach. 4 - Technoeconomic and Policy Considerations for Large-scale Solar Deployment in India Aimee Curtright, Senior Physical Scientist, RAND Corporation, 4570 Fifth Ave, Suite 600, Pittsburgh, PA, 15213, United States of America, acurtrig@rand.org, Zhimin Mao, Oluwatobi Oluwatola, Mridula Dixit Bharadwaj The SERIIUS consortium aims to develop and assess PV and CSP solar technologies that can support India’s ambitious solar deployment goals, recently increased to a target of 100 GW by 2022. RAND and CSTEP are collaborating to conduct technoeconomic and policy analyses to support the SERIIUS consortium. This presentation will discuss recent and ongoing work, including progress made by U.S.-based Pardee RAND students during their visiting internships at CSTEP in India. Engineering Building North, Columbia, MO, 65211, United States of America, mcgarveyr@missouri.edu 1 - Using Big Data to Inform Mental Health Policies

242