2016 INFORMS Annual Meeting Program

POSTER SESSION

INFORMS Nashville – 2016

An Optimization Approach To Detection Of Epistatic Effects Maryam Nikouei Mehr, Graduate Student, Iowa State University, 3004 Black Engineering, Ames, IA, 50011, United States, mnmehr@iastate.edu, Lizhi Wang Epistasis refers to the phenomenon where the interaction of multiple genes affects a certain phenotype more than they do separately. Similar epistatic effects are also ubiquitous in other application areas, where a certain effect is only observable when a particular combination of multiple factors is present. Due to the enormous solution space, it’s hard to detect the epistatic effect. We propose an optimization model that attempts to detect epistatic effects where a large number of observations are available for a relatively small number of explanatory factors. We will share our preliminary results and discuss future research directions. Measuring Competition Between Spanish Engineering Schools Jordi Olivella, Universitat Politecnica de Catalunya, Avda. Diagonal 647, Barcelona, 08860, Spain, jorge.olivella@upc.edu, Fernando Terres In Spain the higher education institution choice is highly affected by the distance between a student’s family home and the institutions. The higher education market is, at least in part, geographically based. Measuring competition among higher education centers needs to take also into account the specializations offered, the number of students admitted and the tuition fees. Several measures are proposed and tested. They are applied to the Spanish Engineering Schools. Ambulance Dispatching Problem To Minimize Response Time And Hospital Congestion Using Approximate Dynamic Programming Seonghyeon Park, Yonsei University, 29, Yonsei-ro 11-gil, 403-ho, Seoul, 03788, Korea, Republic of, s.park10@yonsei.ac.kr Ambulance dispatching problem is to decide which ambulance to send to an emergency call. Previous literature has mainly focused on minimizing response time to an emergency call. However, in the environment where congestions of each emergency room are quite different, it’s important to determine to which hospital to transport patients to treat them efficiently. In this paper, an approximate dynamic programming model is suggested to optimize ambulance dispatching, minimizing response time as well as decreasing hospital congestion. In addition, a case study based on real data is performed to demonstrate the proposed model performs better in comparison with the existing ones. Quantifying The Benefits Of Continuous Replenishment Program For Partner Selection Payam Parsa, PhD Candidate, University of Arkansas, Fayetteville, AR, United States, pparsa@uark.edu, Manuel D Rossetti, Shengfan Zhang, Edward A Pohl Supply chain collaboration programs such as Continuous Replenishment Program (CRP) face challenges with regard to sharing the financial benefits. Supply chain partners often suffer from the ambiguity that exists with the Return on Investment (ROI) of the collaboration programs. This research provides a multi- echelon supply chain model that quantifies the benefits of a continuous replenishment program (CRP) for both partners, and at three levels of inventory holding, transportation and ordering cost component. The model is adopted by a major healthcare manufacturer, with thousands of products and hundreds of demand points, in the form of a software tool. Using An Ontology To Create Content For Clinical Assessment Questions Anna Perini, Innovative Knowledge Representative Specialist, Elsevier, 1600 JFK Blvd, Philadelphia, PA, 19103, United States, a.perini@elsevier.com Using an ontology to create content for clinical assessment questions. This was done by modeling patterns of existing questions and building templates to modify existing questions using our ontological relations to create a new question. Profile Monitoring Using Non-parametric Models For Poisson Data Sepehr Piri, Virginia Commonwealth University, 1015 Floyd Ave., PO Box 842014, Richmond, Richmond, VA, 23284, United States, piris@vcu.edu Profile monitoring is a relatively new technique used to monitor the functional relationship between a response variable and one or more explanatory variables at each time period. Although many studies have been conducted in this field, in most of them, the distribution of the response variable is assumed to be normal which is not always appropriate. To our knowledge, few works have used profile monitoring for poisson data. In this study, we aim to use non-parametric approaches in profile monitoring for those situations where the appropriate distribution is defined by the poisson.

Joint Inventory Replenishment For High Variety Mass Customizers

Michael Prokle, PhD Candidate, University of Massachusetts- Amherst, 290 N Pleasant Street, Apt 2, Amherst, MA, 01002, United States, mprokle@umass.edu, Ana Muriel We address the joint inventory replenishment problem faced by a manufacturer that builds unique products to customer’s specifications. Historic part usage shows lumpy & intermittent demand. The objective is to find a joint part replenishment policy that incorporates the status of the current order pipeline and balances inventory, ordering, and stock-out costs, under given MOQ and lot size requirements. In a case study of a small-size, fast-growing mass customizer, our computational results show that a coordinated part inventory policy results in higher customer service, virtually eliminating lost sales, while lowering cost by taking advantage of shipping economies of scale. Cross Price Elasticities In Retail Price Optimization Jagdish Ramakrishnan, Walmart Labs, San Bruno, CA, United States, jramakrishnan@walmartlabs.com, Mátyás Sustik In store retail, cross price effects have a significant impact on product sales. Determining and estimating cross price elasticities for a large number of products is a challenging problem. We use categorical information and LASSO to estimate a sparse cross item set. We then solve a convexified price optimization problem. Evolving Airplane Boarding Zone Plans Ed Ramsden, Consultant, 1080 County Street, Attleboro, MA, 02703, United States, earamsden@comcast.net To manage the boarding process and reduce boarding times, airlines often assign passengers into a series of ‘boarding zones’. This presentation describes a methods of developing improved improved boarding zone assignment plans through the use of a passenger-level boarding simulation model combined with an evolutionary optimization algorithm. Decision Facing Ambiguity: Mdp, Pomdp And Beyond Mohammad Rasouli, PhD Candidate, University of Michigan, 430 South Fourth Ave, Ann Arbor, MI, 48104, United States, rasouli@umich.edu While most of the decision making tools are developed for a Bayesian framework where the decision maker knows full stochastic description of uncertainties in the environment, decision facing ambiguity (model uncertainty and non-stochastic uncertainty) is a better approach for modeling a lot of practical situations. We discuss how decision making tools including MDP, POMDP, learning (e.g. Multi- armed bandit) and team decision making can be extended for environments with ambiguity. We discuss robustness and bounded rationality in this framework. Optimizing Socioeconomic Balances In Schools Rebecca Reddoch, Furman University, 3300 Poinsett Highway, Greenville, SC, 29613, United States, lattie.reddoch@furman.edu Does the socioeconomic class of a student’s peers matter in the student’s ability to learn? Several studies have suggested that it does. Despite the identification of socioeconomic status as a correlating factor between education and achievement, there are still large performance gaps in high schools throughout the nation. Zoning based on distance ideally provides convenience and minimal travel costs for students, but it is effectively zoning by neighborhood and socioeconomic status. Here we study a multi-criteria model that assigns students to schools based on a combination of socioeconomic and distance factors. Sterilization Network Design Ahmed Saif, Postdoctoral Fellow, HEC Montréal, 3000, Chemin de la Côte-Sainte-Catherine, Montréal, QC, H3T 2A7, Canada, ansaif1976@yahoo.com Centralizing sterilization services in hospital networks can cut cost and improve efficiency through better utilization of resources, risk-pooling and economies-of- scale. We compare three organization schemes: fully distributed, centralized processing, and centralized processing and stock keeping. The sterilization network design problem is formulated as a mixed-integer concave minimization program, then reformulated as a mixed-integer second-order cone program with a piecewise-linear cost function so it can be solved efficiently. Testing is done on a realistic case study under different scenarios. The cost components in every scheme are analyzed and managerial insights are drawn. An Integrated Facility Location And Network Restoration Model Under Repair Time Uncertainty Ece Sanci, PhD Pre-Candidate Student, University of Michigan, 1205 Beal Avenue, Ann Arbor, MI, 48109, United States, ecesanci@umich.edu, Mark Stephen Daskin We propose a two-stage stochastic programming model for an integrated facility location and network restoration problem in a disaster-prone region where facility location decisions should be made in the pre-disaster stage. We capture uncertainty in the network availability by incorporating the repair times required to restore the damaged arcs. In contrast to other models that ignore repair times, our model locates some facilities in remote, low-demand areas that are unreachable for a certain number of periods following a disaster.

295