Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SD76

4 - Capability Analysis for Canada’s Special Operations Forces Ramzi Mirshak, DRDC, Centre for Operational Research and Analysis, Ottawa, ON, Canada The Canadian Special Operations Forces Command (CANSOFCOM) is developing its future operating concept. To guide this evolution, wargames were used to review capabilities of key missions both now and in the future. Assessments included ordinal scoring of criteria and gathering of comments. We present the methods and approach used, including efforts to minimize respondent bias. n SD76 West Bldg 212C New Developments in Planning Production and Inventories General Session Chair: Reha Uzsoy, North Carolina State University, Raleigh, NC, 27695-7906, United States 1 - Planning Releases during Product Transitions in Semiconductor Manufacturing Atchyuta B. Manda, North Carolina State University, Campus Box 7906, Raleigh, NC, 27695-7906, United States, Reha Uzsoy We use simulation optimization to model the introduction of a new product into an operating semiconductor manufacturing facility, which can have significant adverse effects on both current and new products. Our results show that planning releases by simulation optimization model yields significant improvements in performance over simple alternatives, and provides intuitive solutions. 2 - A Matheuristic for Integrated Production Planning and Order Acceptance Decisions Lars Moench, Professor, University of Hagen, Universitaetsstrasse 1, Hagen, 58097, Germany, Hung-Kai Wang We discuss a production planning model with load-dependent lead times. Order acceptance decisions with respect to flexible due dates are allowed in addition to release decisions. We show that the resulting production planning problem is NP- hard. A variable neighborhood search (VNS) technique is hybridized with linear programming to solve large-sized problem instances. Results of computational experiments for problem instances that are derived from a scaled-down wafer fab model are presented. The results demonstrate that the proposed matheuristic outperforms time-based decomposition approaches from the literature 3 - Optimizing End-item Inventory Control under Rolling Scheduling A.G. de Kok, Eindhoven University of Technology, Eindhoven, Netherlands We discuss a simulation-based methodology to optimize end-item inventories in multi-item multi-echelon inventory systems. We build on a sufficient condition for the Newsvendor equation to hold for end-items. This condition requires that a change in an end-item control parameter yields the same change in the net stock of this end-item at each point in time. This sample path property can be exploited for an efficient simulation-based approach to find the optimal end-item inventory control parameters. 4 - Extending Clearing Functions by Including the Process History: Approaches and Research Gaps Hubert Missbauer, University of Innsbruck, Universitaetsstrasse 15, Innsbruck, 6020, Austria The fit of clearing functions to empirical or simulated data can be improved substantially by including explanatory variables that reflect the history of the process prior to the period under consideration. We show by means of numerical examples that the resulting multi-dimensional clearing functions cannot be integrated into existing clearing function models in a straightforward manner. Using simulation results we eleborate on the discrepancy between clearing function fit and performance of the order release model. These results, together with some structual properties of one-dimensional clearing function models, raise Ivo Adan, Eindhoven University of Technology, Den Dolech 2, Eindhoven, 5600 MB, Netherlands, Jelle Adan, Stephan Sneijders, Alp Akcay We present a reenactment fluid-flow simulation-based method to optimize buffer sizes in a semi-conductor assembly line. The optimization method is a combination of a discrete-event fluid-flow simulation model and a multi-start search algorithm. The method is applied to a real-world case in semiconductor manufacturing. the question what clearing functions really are. 5 - Reenactment Fluid Flow Simulation in Semiconductor Manufacturing

n SD77 West Bldg 213A Disaster Operations Management Sponsored: Public Sector OR Sponsored Session Chair: Christopher Zobel, Virginia Tech, Virginia Tech, Blacksburg, VA, 24061-0235, United States Co-Chair: Andrew N. Arnette, University of Wyoming, Laramie, WY, 82071-2000, United States 1 - Extensions to Pre-Positioning Disaster Relief Asset Optimization Model Andrew N. Arnette, University of Wyoming, 1000 E. University Avenue, Dept 3275, Laramie, WY, 82071-2000, United States, Christopher Zobel In this work, we extend an existing model for disaster relief asset pre-positioning so that it applies to a range of different types of resources, some of which may be perishable or substitutable. The relative performance of the general model formulation is examined with respect to problems such as allocating medical supplies to support post-disaster relief efforts. 2 - Forecasting for Disaster Management Nezih Altay, DePaul University, 1 E. Jackson Blvd., Driehaus College of Business, Chicago, IL, 60604, United States We present the results of a structured literature review on the use of forecasting methods in for disaster management. Based on our findings we identify research gaps and make recommendations for future research. 3 - Modeling Cascading Failures and Recovery of Interconnected Systems: Case Study of Puerto Rico after Hurricane Maria Felipe Aros-Vera, Ohio University, 277 Stocker Center, 1 Ohio University, Athens, OH, 45701, United States, Shital Thekdi Understanding interdependencies and cascading effects is of key importance for the design of critical infrastructure and further response and recovery from disruptions. This work studies these interdependencies using network design and risk-management tools. The presentation includes empirical evidence and a case study of interdependent infrastructure systems after Hurricane Maria in Puerto Rico. 4 - Ilicit Drug Abuse and Seizure Patterns: Insights from Data Hossein Najmi, University of North Texas, Denton, TX, United States, Shailesh S. Kulkarni, Debjit Roy Approximately 90 people die every day in the USA as a result of illicit drug abuse and in 2017, President Trump declared the opioid crisis a health emergency. However, there are few studies on the patterns of drug seizure, especially using advanced analytics. To address this limitation, we collected data on all drug seizures between 2010-2016 at the state level, classified them based on the Drug Schedule of the Drug Enforcement Administration (DEA) and performed data analysis on over hundred thousand data points to gain insight into the dynamics of drug seizure. By studying patterns, this study tries to develop a potential tool to increase seizure amounts and to help better control the drug disaster. n SD78 West Bldg 213B Location Models II Sponsored: Location Analysis Sponsored Session Chair: Zvi Drezner, California State University-Fullerton, Fullerton, CA, 92834, United States 1 - A Cover Based Competitive Facility Location Model with Continuous Demand Zvi Drezner, California State University Fullerton, Steven G. Mihaylo College of Business and Economics, Dept of ISDS, Fullerton, CA, 92834, United States, Tammy Drezner, Atsuo Suzuki We propose and solve a competitive facility location model when demand is continuously distributed in an area and each facility attracts customers within a given distance. This distance is a measure of the facility’s attractiveness level which may be different for different facilities. The market share captured by each facility is calculated by two numerical integration methods. These numerical approaches can be used for evaluating functional values in other Operations Research models.

118