2016 INFORMS Annual Meeting Program

MA44

INFORMS Nashville – 2016

2 - Agent-based Modeling For Resilience Of Disaster Recovery Fei He, Texas A&M University, Kingsville, Kingsville, TX, United States, Fei.He@tamuk.edu, KumareshBabu Murugesan Disaster recovery involves multiple stakeholders and various uncertainties associated with environment and community behaviors. This research use agent- based modeling and game theory to investigate the uncertainty and interdependence in the household and business for disaster recovery. The effects of disaster mitigation, and stake holders’ learning capability to disaster recovery are investigated. 3 - Modeling Fire Risk And Resource Allocation For Fire Protection And Safety Vineet Madasseri Payyappalli, PhD Student, University at Buffalo, SUNY, Buffalo, NY, United States, vineetma@buffalo.edu, Adam Behrendt, Jun Zhuang Fire-related hazards are an everyday phenomenon, and firefighting in the United States owe to more than one million firefighters in about 30,000 fire departments across the country. The estimated total cost of fire was $329 billion in 2011, and yet there is little work in the literature about risk assessment, cost-benefit analysis, and resource allocation in fire protection. Using a data-driven study, we propose empirical and theoretical models to assess risk levels and develop risk- reduction strategies that include optimal resource allocation, optimal facility design, and optimal routing solutions. 4 - Crisis Information Distribution Among Official Users In Twitter Based On Hurricane Sandy Bairong Wang, University at Buffalo, The State University of New York, 338 Bell Hall, Buffalo, NY, 14260, United States, bairongw@buffalo.edu, Jun Zhuang Our study analyzes how crisis information about Hurricane Sandy is distributed from official Twitter accounts to the common Twitter users based on 6 social media key performance indicators (KPIs). Our results show that (a) the six KPIs are significantly different among governmental organizations (GOs), non- governmental organizations (NGOs) and news agent users; (b) the networks formed by mention and re-tweet are effective methods to reach more public members; (c) the information coverage could be expanded to more stakeholders in disaster scenario; (d) distributing speed is faster among networks formed by news agent users than GO and NGO users. MA44 208B-MCC Environmental Decision Analysis Sponsored: Decision Analysis Sponsored Session Chair: Melissa A Kenney, University of Maryland, College Park, MD, United States, kenney@umd.edu 1 - Robust Decision Making Methods For Water Resource Management Under Climate Change Uncertainty Seth Guikema, University of Michigan, sguikema@umich.edu, Julie Shortridge, Ben D Zaitchik Water resource systems must be managed under deep, long-run uncertainty about the impacts of climate change on water quantity and quality in a given basin. Standard approaches for decision making under uncertainty have limitations in this context. We summarize work done to further develop an alternative approach, based on the Robust Decision Making method (RDM). RDM seeks not to find the optimal solution to a given management problem but to find solutions that are robust in the sense of doing well under a wide range of future conditions. We demonstrate the method with an application to water resource management in the Lake Tana basin. 2 - Multi-criteria, Interactive Optimization For Design Of Watershed Plans

3 - Using Visualization Science To Diagnose And Improve Global Change Indicator Understandability Michael D Gerst, University of Maryland, mgerst@umd.edu Melissa A Kenney Indicators are variables that stakeholders believe summarize relevant trends. They have become an increasingly important part of continuous assessment of global environmental change. For indicators to be effective, they need to be understood by diverse audiences. Using visualization science, we have diagnosed and redesigned a set of global change indicators, showing how simple visual changes can lead to large improvements in understandability. 4 - Understanding Homeowner Decisions On System Configuration A common problem faced by many older U.S. cities is the management of storm water to prevent the over flow of sewage and pollutants into local waterways. Many cities have addressed this problem through grey infrastructure improvements, as well as promoting low impact development (LID) on private and public property. Previous researchers have proposed models for assessing optimal LID installations and predicting consumer behavior in response to incentives. We explore what an optimal storm water management program would look like in light of these models, with the high level aim of learning best practices in system configuration from observing consumer choices. MA45 209A-MCC Agent-based Modeling and Simulation - Overview Sponsored: Simulation Sponsored Session Chair: Charles M Macal, Argonne National Laboratory, Argonne, IL, United States, macal@anl.gov 1 - Agent-based Modeling And Simulation – Overview Charles M Macal, Argonne National Laboratory, macal@anl.gov Agent-based modeling and simulation (ABMS) is an approach to modeling systems comprised of autonomous, interacting agents. Applications are growing rapidly in fields ranging from modeling the stock market to predicting the spread of epidemics. Complex adaptive systems, emergent behavior, and self- organization are a few of the notions from ABMS. This session provides an overview of ABMS, covers its foundations, development toolkits and methods, practical aspects, and the relationship of ABMS to conventional OR. Key ABMS resources, publications, and communities are identified. It concludes by suggesting research challenges to advance the field of ABMS for the coming years. MA46 209B-MCC Pricing, Revenue Management and Operations in Retailing Sponsored: Revenue Management & Pricing Sponsored Session Chair: Mehmet Sekip Altug, Assistant Professor, George Washington University, 2201 G. Street, NW, Washington, DC, 20052, United States, maltug@gwu.edu 1 - Assortment Optimization Under A Synergistic Version Of The Multinomial Logit Model Venus Lo, Cornell University, Ithaca, NY, United States, vhl8@cornell.edu, Huseyin Topaloglu We consider the revenue management problem of offering an optimal subset of goods when there are product synergies. The traditional multinomial logit choice model suffers from Independence of Irrelevant Alternatives and offering a larger subset must decrease each goods’ choice probabilities. Our synergistic model has a similar structure but offering selected pairs of goods together can boost their choice probabilities. In the special case where synergy exists in a linear fashion, we provide an efficient dynamic program and show that the optimal subset can be found in one step by solving a simple linear program. For Parcel Level Storm Water Management Royce Francis, George Washington University, seed@email.gwu.edu, Domenico C Amodeo

Andrew Hoblitzell, Indiana University Purdue University Indianapolis, Indianapolis, IN, 46202, United States, ahoblitz@umail.iu.edu, Meghna Babbar-Sebens, Snehasis Mukhopadhyay

Multi-objective optimization has yielded numerous algorithms for design of solutions to real-world planning problems. The inclusion of decision makers (DMs) within the optimization algorithm’s search process, especially for planning problems with DM-specific, subjective, qualitative, and/or unquantifiable criteria, has gained interest recently. Our work focuses on modifications made to our existing watershed planning decision support system, called WRESTORE. Interactive genetic algorithms and reinforcement-based machine learning algorithms are used for search and optimization, while neural networks and other methods are utilized for the modeling of human users’ criteria.

136