Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MB77

n MB77 West Bldg 213A Disaster Preparedness, Response, and Recovery Sponsored: Public Sector OR Sponsored Session Chair: Pinar Keskinocak, Georgia Institute of Technology, Georgia Institute of Technology, Atlanta, GA, 30332, United States Co-Chair: Seyma Guven Kocak, Georgia Institute of Technology, Atlanta, GA 1 - Inequity-averse Shelter Location for Disaster Preparedness Sibel Salman, Koc University, Rumeli Feneri Yolu, Istanbul, 34450, Turkey, Mahdi Mostajabdaveh, Walter J. Gutjahr We address selecting capacitated shelter locations in preparation for disasters under demand uncertainty and random transportation network disruptions by an inequity-averse approach. We develop a 2-stage stochastic programming model to minimize a linear combination of efficiency and inequity measures. Moreover, a service level is imposed by a chance constraint. We develop a tailored genetic algorithm (GA) to solve this problem for large real-life and benchmark instances. Our computational results show that the GA outperforms Cplex in terms of run time and solution quality. We derive insights from a case involving an earthquake-prone district of Istanbul. 2 - The Post-disaster Debris Clearance Problem with Resource Allocation and Learning In the aftermath of a disaster, debris needs to be cleared quickly and efficiently in order to continue critical response operations. Motivated by the debris clearance operations, this work addresses a problem of establishing connectivity on a disrupted network with inherent uncertainty and learning. Since the information on the clearance times of the roads only becomes available after the network is explored, this problem possesses exploitation / exploration trade-off. Considering learning and exploration / exploitation trade-off, we develop various exact and approximate solution methods in order to optimize the benefit accrued over time by connecting critical points in the network. 3 - Time Synchronization Issues in Emergency Response Jordan Srour, Assistan Professor of Operations Management, Lebanese American University, Adnan Kassar School of Business, P.O. Box 13-5053, Beirut, 1102 2801, Lebanon In any city - smart or not - emergency response is a critical service. In smart cities, the use of technology to manage access to and dispatching of emergency vehicles is particularly important. However, when a system must manage processes spanning multiple computers, clock drift becomes a prominent issue. In a case study with an existing emergency response provider, we show how analytics can be used to determine the impact and source of clock drift. We also show the impact that system design techniques can have in remedying this issue. 4 - A Stochastic Integer Programming Framework for Large Scale Patient Evacuation under a Forecasted Disaster Kyoung Eric Kim, University of Texas-Austin, Austin, TX, United States, Erhan Kutanoglu, John Hasenbein A major humanitarian challenge of a forecasted disaster such as a hurricane is to decide whether or not the patients in hospitals and nursing homes in the affected area should be evacuated. We propose a comprehensive modeling and methodological framework for large scale patient evacuation when an area is faced with such a disaster. We integrate a scenario generation scheme that uses updated hurricane forecasts and a scenario-based stochastic integer program to make decisions on resource allocation (for staging area locations, ambulances/ambuses, and sending and receiving staff) and patient movements. Real-world data from Southeast Texas region is used in our computational study. 5 - Scheduling Staff for the FEMA National Response Coordination Center Erica L. Gralla, George Washington University, 800 22nd st NW, Rm 2680, Washington, DC, 20052, United States, Kai Friesecke, Shelby Grumer, Jillian D’Arrigo, Hernan Abeledo, Joseph Barbera The Federal Emergency Management Agency’s National Response Coordination Center coordinates federal support for response to major disasters. Producing staff schedules for the center is a challenge, since most staff have other commitments, and each position requires specific qualifications. We describe a scheduling approach that efficiently rosters qualified personnel and minimizes the violation of staff preferences, including days off, manager preferences, and others. Results show which preferences are hard to satisfy and provide recommended roster sizes and cross-training requirements. Seyma Guven-Kocak, Georgia Institute of Technology, 5211 Denmeade Dr, Atlanta, GA, 30345, United States, Pinar Keskinocak

n MB78 West Bldg 213B Joint Session SOLA/PSOR/Practice Curated: Public Sector Location Applications Sponsored: Location Analysis Sponsored Session Chair: Rajan Batta, University at Buffalo (SUNY), Buffalo, NY, 14260, United States 1 - Prepositioning of Supplies in Domain of Natural Disasters Logistics: A Systematic Review of Operations Research and Management Science Papers Monir Sabbaghtorkan, University at Buffalo (SUNY), Amherst, NH, United States, Rajan Batta, Qing He Prepositioning of supplies is an efficient way to improve the post-disaster relief efforts. In this paper, we explore and review OR/MS papers related to this topic to find the gaps and give directions to future researchers in this area. The classifications that are used in this paper are as follow, Authors affiliation, journals contribution, number of papers over the years and type of disaster considered by papers. Moreover, we would present the mathematical model features for each paper, like objective function, decision variables, constraints, solution methodology and etc. 2 - Responding to Medical and Fire Emergencies on a Disrupted Network Gina Galindo, Universidad del Norte, KM5 Antigua via a Puerto Colombia, Puerto Colombia, Colombia, Juliette Garc a Events such as heavy rains or extreme snowfall may disrupt roads and affect the response times of critical services such as fire and medical emergencies. This research develops a methodology for effectively allocating medical and fire emergencies to available facilities on a disrupted network. The methodology was applied to a case study in Barranquilla, Colombia. Based on the results, we recommend potential site for locating additional fire and medical facilities. 3 - A Stochastic Programming Model for Emergency Supply Planning Considering Traffic Congestion Xiaofeng Nie, Texas A&M University, College Station, TX, United States, Qingyi Wang Traffic congestion delays emergency supply after disasters, but it is seldom considered in the emergency logistics literature. To fill this gap, we propose a traffic congestion incorporated two-stage stochastic programming model that facilitates the planning of supplies pre-positioning and post-disaster transportation. The corresponding mixed-integer nonlinear program is efficiently solved with a generalized Benders decomposition algorithm. A case study on a hurricane threat in the southeastern U.S. shows the superiority of our model and provides managerial insights. 4 - A Bi-Level Combined Facility Location and Network Design Problem in Hazardous Materials Transportation Xufei Liu, University of South Florida, Tampa, FL, United States, Changhyun Kwon In this paper, we consider a bi-level optimization problem for hazardous materials transportation to determine facility locations and network design. The leader intends to minimize the total cost of facility costs and exposure risk by restricting the shipment and anticipates the reaction of the followers who want to minimize the transportation costs by choosing the shortest path for shipments. Considering the uncertainty of the number of trucks and exposure risk, we formulate a robust bi-level combined facility location-network design model and propose a cutting plane algorithm.

176