Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TC41

Disruptive Events Cameron MacKenzie, Iowa State University, 3029 Black Engineering, Industrial and Manufacturing Systems Eng, Ames, IA, 50011, United States, Brandon Landowski

North Bldg 226A Probability Models in a Shared Economy Sponsored: Applied Probability Sponsored Session Chair: Mariana Olvera-Cravioto, University of California-Berkeley, Berkeley, CA, 94720, United States 1 - Last-mile Shared Delivery: A Discrete Sequential Packing Approach Junyu Cao, University of California-Berkeley, Berkeley, CA, 94720, United States, Mariana Olvera-Cravioto, Max Shen, Max Shen We propose a model for optimizing the last-mile delivery of n packages from a distribution center to their final recipients, using a strategy that combines the use of ride-sharing platforms (e.g. Uber or Lyft) with a traditional van infrastructure. Our technical approach is based on the formulation of a discrete sequential packing problem, which is closely related to Renyi’s parking problem. The main objective is to compute the optimal reward offered to private drivers for each individual package in a way that minimizes the total expected cost for delivering all packages. We show that under natural assumptions our mixed strategy can achieve significant improvements compared to a van-only strategy. 2 - Taxi-customer Queues with Delayed Matching Lu Wang, Univeristy of North Carolina at Chapel Hill, B52 Hanes Hall, Chapel Hill, NC, 27599, United States, Vidyadhar Kulkarni We consider a network of taxis and customers with Poisson arrivals and exponential patience times. We assume that there are exponentially distributed delays in matching the taxis with customers. We establish two methods to compute the fluid and diffusion approximations for the queue lengths and compare their performance. 3 - Customer Preference and Station Network in the London Bike Share System Pu He, Columbia University, Uris Hall, Cub 4H, New York, NY, 10027, United States, Fanyin Zheng, Elena Belavina, Karan Girotra We study customer preference for the bike share system in the city of London. We estimate a structural demand model on the station network to learn the preference parameters and use the estimated model to provide insights on the design and expansion of the bike share system. We highlight the importance of network effects in understanding customer demand and evaluating expansion strategies of transportation networks. We develop a new method to deal with the endogeneity problem of the choice set in estimating demand for network products. Our method can be applied to other settings, in which the available set of products or services depends on demand. 4 - Managing Services with Dependent Valuations and Service times Achal Bassamboo, Ohad Perry, Chenguang Wu The valuation for service of an arriving customer often depends on his individual service requirement; in this work we consider a queueing model in which these two random variables are stochastically dependent. Specifically, customers are price and delay sensitive, and decide whether to queue for service based on their service valuations, waiting cost and the price of service. Employing a general dependence order, we show that the provider’s optimal revenue decreases with the strength of the dependence. Moreover, considering the valuation and service requirement to be independent when they are in fact dependent can lead to substantial revenue losses. n TC40 North Bldg 226B Risk and Decision Analysis for the Resilience of Infrastructure Systems and Communities Sponsored: Decision Analysis Sponsored Session Chair: Hiba Baroud, Vanderbilt University, Nashville, TN, 37235, United States 1 - Modeling Risks of Extreme Weather Induced Power Outages Sayanti Mukherjee, Assistant Professor, University at Buffalo, Buffalo, NY, 14260, United States Weather extremes cause sustained power outages that affect millions of people, causing extensive economic losses every year. The importance of building electricity sector’s resilience, and thereby enhancing its service-security and long- term economic benefits is well established. To characterize the key predictors of such power outages, we propose a multihazard risk estimation approach. Our model suggests that power outage risk is a function of hazard types, expanse of overhead power systems, expanse of rural or urban areas, and levels of investment in operations/maintenance activities. The proposed model can help utility regulators to make risk-informed resilience investment decisions. 2 - A Multiple Decision-maker Approach to Allocating Resources for

Most decision-making models for disruptive events and resilience consider a single decision maker. However, building community resilience and preparing and responding to disruptive events involve multiple stakeholders and different decision makers. We create different resource allocation models for different decision makers preparing and responding to a disruptive event. The decision makers include the federal government, state government, private sector, and non-governmental organization. We explore how the benefits of cooperating among these decision makers can increase the effectiveness of allocating resources to prepare and respond to disruptive events. 3 - Climate Analysis Support Tool for Community Resilience Assessment & Adaptation Planning Leslie Gillespie-Marthaler, Vanderbilt University, 512 Cedar Cove, Nashville, TN, 37209, United States, Hiba Baroud Communities require methods to aid in use of data related to climate change, as well as means to apply scenario-based alternatives to current and future decision making processes. To this end, we have developed a scalable tool using downscaled CMIP5 multi-model ensemble outputs, Rstudio, and local data to assess how extreme events may change in the future. The tool can be used to develop possible scenarios related to changes in frequency and magnitude for extreme precipitation (100, 200, 500 year) events. Outputs from the tool will be used in assessing community resilience to flood for a test community to build strategies that enhance current and future survival, wellbeing, and sustainable resilience. 4 - Optimal Inspection Strategies for Turbine Runners in Hydropower Plants Andreas Kleiven, PhD Student, Norwegian University of Science and Technology, Alfred Getz vei 3, Trondheim, 7491, Norway, Stein-Erik Fleten, Carl Ullrich, Benjamin Fram We develop a model for stochastically deteriorating turbine runners in hydropower plants with the purpose of finding the implied cost of inspections. Unscheduled downtime caused by failure of critical components in a hydropower plant may cause huge revenue losses. Inspections give information about the system state, and maintenance actions can be planned accordingly. In this work we simulate and analyze an inspection and replacement strategy based on turbine data from Norwegian hydropower plants. Participatory Modeling Sponsored: Decision Analysis Sponsored Session Chair: Karen Jenni, U.S. Geological Survey, DFC, MS 939, Denver, CO, 80225, United States Co-Chair: Antonie Jetter, Portland State Universisty, Portland, OR, United States 1 - Tools and Methods in Participatory Modeling Karen Jenni, U.S. Geological Survey, DFC, MS 939, Denver, CO, 80225, United States, Alexey Voinov, Steven Gray, Pierre Glynn, National SESYNC Participatory Modeling Working Group The diversity of tools and methods used in participatory modeling can create challenges for stakeholders and modelers when selecting the ones most appropriate for their projects. We offer a systematic overview, assessment, and categorization of methods to assist modelers and stakeholders with their choices and decisions. Based on the results of our Participatory Modeling Working Group, and a survey of modelers engaged in participatory processes, we offer practical guidelines to improve decisions about method selection at different stages of the participatory modeling process. 2 - Participatory Exploratory Modelling and Analysis with Fuzzy Cognitive Maps Antonie J. Jetter, Portland State University, ETM Department, 1900 SW 4th Ave, Portland, OR, 97207, United States Participatory modelling pools participants’ partial system knowledge, often resulting in “deep uncertainty because some aspects of the system structure remain unknown. Exploratory modeling and analysis (EMA) simulates system behavior for multiple unknown inputs and alternative system structures. We illustrate a novel approach to combine EMA with Fuzzy Cognitive Mapping with two studies, which used different method to determine the range of “deep uncertaintyö, namely content analysis and expert surveys. We present data collection and EMA approaches, computational challenges, results, and lessons learned for these strategies. 3 - A Participatory Model for Improving Safety Culture in Oil and Gas Operations n TC41 North Bldg 226C

321