2016 INFORMS Annual Meeting Program

SA32

INFORMS Nashville – 2016

4 - Dual Sourcing: Optimal Procurement Policy With Option Hedging Against Freight Rate Risk Arun Chockalingam, Eindhoven University of Technology, Hoog Gagel 62, Eindhoven, 5611BG, Netherlands, a.chockalingam@tue.nl, Taimaz Soltani, Jan C Fransoo, Chung-Yee Lee Raw materials cost less to procure on offshore markets as opposed to domestic markets. However, offshore procurement requires ocean shipping, the price of which is highly volatile, and firms usually have to charter ships, even if they do not use the whole capacity of ships. We consider a commodity processor with two sourcing choices and develop models to determine the firm’s optimal sourcing policy. The models allow for three sources of uncertainty: demand, commodity spot price and freight rate. Using option contracts as hedging tools against freight rate risk, we develop a model that integrates a firm’s optimal sourcing decision with the integer constraint on hiring ships. SA32 203A-MCC Scheduling I Contributed Session Chair: Neil Desnoyers, Saint Joseph’s University, 133 Green Valley Rd, Upper Darby, PA, 19082, United States, ndesnoye@sju.edu 1 - A Bicriteria Scheduling Problem With Two Competing Agents Single machine scheduling with two competing agents is studied. Cost function of the machine and agent 1 must be minimized simultaneously, while cost function of agent 2 is keeped below a determined value Q. Total completion time and maximum lateness of the machine, and most regular cost functions of the agents are consided. The model of multiagent scheduling has been enlarged to including the goal of both machine and agents. Some polynomial and pseudo-polynomial algorithms are presented. 2 - Scheduling Student Volunteers For The Informs Annual Meeting Neil Desnoyers, Adjunct Instructor, Saint Joseph’s University, 5600 City Avenue, Philadelphia, PA, 19131, United States, ndesnoye@sju.edu The 2015 INFORMS Annual Meeting in Philadelphia required the assistance of 59 student volunteers. Each student volunteer was required to serve two half-day shifts out of the eight shifts available over four days (AM and PM shift each day). Via a web survey, student volunteers indicated five of eight shifts they were available to work. Between 12 and 14 student volunteers were required for each shift, and student volunteers were assigned to 9 or 10 locations/roles each shift. The problem was set up and solved as a binary integer programming problem. The problem provides lessons for volunteer scheduling in general. SA33 203B-MCC Simulation and Optimization I Contributed Session Chair: Tomas Ignacio Lagos, Masters Degree Student, University of Chile, 2017 - Pozuelo, Santiago, 7640031, Chile, tomas.lagos.gonzalez@gmail.com 1 - An Integrated Multi-criteria Decision Making And Simulation- optimization Framework For Supplier Selection Shudong Sun, Northwestern Polytechnical University, 127 Youyi West Road, PO Box 554, Xi’an, 710072, China, sdsun@nwpu.edu.cn Mohammad Dehghanimohammadabadi, Teaching Assistant Professor, Northeastern University, 360 Huntington Ave, MIE Department, Boston, MA, 02115, United States, mdehghani@neu.edu, Emanuel Melachrinoudis A wide spectrum of criteria have been introduced by researchers to evaluate the suppliers’ performance; however, measuring and employment of all of these criteria is impractical. In this study, a two-fold integration of MCDM and Simulation-Optimization is developed to select the most effective criteria for the supplier selection process. In this framework, the MCDM module incorporates a combination of criteria to select the suppliers. Then, the simulation model evaluates the performance of the Supply Chain (SC) considering the selected suppliers. Based on the simulation results, a metaheuristic algorithm finds the best/good combination of the criteria that maximizes the SC performance.

2 - Cloud-based Collaborative Application Development Susanne Heipcke, Principal Engineer, FICO, FICO House, Starley Way, Birmingham, B37 7GN, United Kingdom, susanneheipcke@fico.com, Oliver Bastert Integrated development environments are critical for efficient development but tooling for algebraic modeling languages has been lacking adoption of the latest technologies, no single tool covering the whole application development process so far. FICO Optimization Designer provides a novel approach for collaborative web-based development of optimization solutions. It supports model development in Xpress-Mosel, add-in predictive models implemented in R, deployment as on premises or cloud applications, and debugging of models run from a FICO Optimization Modeler application GUI. 3 - Adaptive Sampling Trust-region Optimization For Derivative- based And Derivative-free Simulation Optimization Problems Sara Shashaani, PhD, Purdue University, Lafayette, IN, 47901, United States, sshashaa@purdue.edu, Raghu Pasupathy We present ASTRO and ASTRO-DF – adaptive sampling trust-region optimization algorithms – or solving derivative-based and derivative-free continuous simulation optimization problems. Sampling in ASTRO and ASTRO-DF is done adaptively in an attempt to keep stochastic and structural errors in lock-step as the algorithm iterates evolve through the search space. We show consistency and discuss finite-time performance for a set of low to moderate dimensional optimization problems. 4 - Designing Resilient Electric Networks Under Natural Hazards Tomas Ignacio Lagos, Masters Degree Student, University of Chile, 2017 - Pozuelo, Santiago, 7640031, Chile, tomas.lagos.gonzalez@gmail.com We present an optimization framework for the problem of designing a resilient electric grid under high impact and low probability events, such as earthquakes. We use an Optimization via Simulation approach to solve this discrete decision problem, where the measure of resilience is the expected Energy Not Supplied and its evaluation uses an existing simulator with historical earthquake data, information of fragility curves of the components provided by FEMA, and an Optimal Power Flow model. We use this framework to evaluate the effect of different resilience measures and algorithms. Chair: Carolyn Queenan, University of South Carolina, 1014 Greene St, Columbia, SC, 29208, United States, carrie.queenan@moore.sc.edu 1 - The Impact Of Call Rotations And Geographic Localization Of Patients On Hospital Performance Douglas Morrice, University of Texas-Austin, Austin, TX, United States, douglas.morrice@mccombs.utexas.edu, Ying Chen, Jonathan F Bard, Luci Leykum We study the impact of an on-call rotation of teaching teams for admissions and geographic localization of patients on hospital performance using patient-level data from a Texas teaching hospital and simulation. Performance is measured by length of stay in the Emergency Department, patient hand-offs, and bed availability. The results of this study inform admission decision-making including patient allocation to medical teams and admission capacity planning. 2 - Complementarities Or Substitutes Of Physician Employment For Managing Patient Care: Effects Of Focus, Experience, And Technology David Zepeda, Northeastern University, Boston, MA, 02115, United States, d.zepeda@neu.edu, Gilbert N. Nyaga, Gary J. Young A dramatic change in the health care industry is the increasing emphasis on linking provider payment to clinical quality performance metrics. This is one of several considerations leading hospitals to vertically integrate by acquiring physician practices and employing physicians directly. Yet, there is little evidence regarding whether this form of vertical integration leads to better performance on clinical quality performance metrics. We empirically evaluate the relationship between hospital employment of physicians and hospital performance on clinical quality performance metrics. We also consider several hospital operational considerations as potential moderators. SA34 204-MCC Hospital Operations Sponsored: Manufacturing & Service Oper Mgmt, Healthcare Operations Sponsored Session

28