2015 Informs Annual Meeting

MC38

INFORMS Philadelphia – 2015

MC36 36-Room 413, Marriott Community-Based Operations Research II Sponsor: Public Sector OR Sponsored Session Chair: Michael P. Johnson, Associate Professor, University of Massachusetts Boston, 100 Morrissey Blvd., McCormack Hall Room 3- 428A, Boston, MA, 02125-3393, United States of America, Michael.Johnson@umb.edu 1 - Resource Allocation for Sustaining Interventions in the Education System Donna Llewellyn, Executive Director, Boise State University, I910 University Drive, ISDI - ACS 104, Boise, ID, 83725-1155, United States of America, donnallewellyn@boisestate.edu, Pratik Mital, Roxanne Moore In this work, the Education System Intervention Modeling Framework (ESIM) is developed that can be used to analyze interventions in the K-12 education system. The framework aids in allocating resources to the more important parts of the system such that probability of sustaining the intervention can be maximized and the cost of implementation remains within the budget constraints. The framework can also be extended to analyze other complex systems like Healthcare, Humanitarian aid etc. 2 - Blending Systems Thinking Approaches for Organisational Diagnosis: Child Protection in England David Lane, Henley Business School, Reading, United Kingdom d.c.lane@henley.ac.uk The Department for Education’s high-profile ‘Munro Review of Child Protection’ used a blend of systems thinking ideas. First, a compliance culture that had emerged was diagnosed. Then system dynamics generated a complex map of the intended and unintended consequences of previous policies and helped identify the sector’s drivers. This led to recommendations that were systemically coherent, avoiding problems produced by previous policies. Government supported and implemented the recommendations. 3 - Multiple Resource Type Straddling a Standard with Applications in Election Resource Allocation Theodore Allen, Associate Professor, The Ohio State University, 1971 Neil Avenue, 210 Baker Systems, Columbus, OH, 43221, United States of America, allen.515@osu.edu, Muer Yang The challenge of guaranteeing that no one will wait over 30 minutes using simulation optimization is explored. Novel selection and ranking methods are proposed. Numerical results illustrate potential new guidelines and associated computational savings. 4 - Measures and Inference of Spatial Access to Pediatric Dental Care in Georgia Monica Gentili, Georgia Tech, North Ave NW, Atlanta, GA, United States of America, mgentili3@mail.gatech.edu, Shanshan Cao, Nicoleta Serban, Susan Griffin We develop a measurement and modeling framework to infer the impact of policy changes on disparities in spatial accessibility to pediatric dental care in Georgia. Our measurement models are based on optimization models that match need of service with supply under a series of user and provider system constraints. We compare the derived measures and evaluate the impact of policy interventions for two population groups (publicly insured and privately insured children) and for rural and urban areas. MC37 37-Room 414, Marriott Health Care Modeling and Optimization VII Contributed Session Chair: Kamil Ciftci, PhD Candidate, Lehigh University, H.S. Mohler Laboratory, 200 West Packer Ave., Bethlehem, PA, 18015, United States of America, kac208@lehigh.edu 1 - A Multi-Objective Algorithm for Optimizing Service Consistency in Periodic Vehicle Routing Problems Kunlei Lian, University of Arkansas, Bell 4113, 1 University of Arkansas, Fayetteville, AR, 72701, United States of America, klian@uark.edu, Ashlea Milburn, Ronald Rardin This research concerns optimizing service consistency in periodic vehicle routing problem, in which customers require repeatable visits over a time horizon and visits to a customer can only happen on one of his allowable visit day combination. Service consistency, including driver consistency and time consistency is optimized together with travel cost using a heuristic multi-objective algorithm. Large neighborhood search is used in the algorithm framework to optimize each objective separately.

2 - Conjugate Gradient Algorithms to Optimize RBE-weighted Dose in Intensity Modulated Proton Therapy Guven Kaya, PhD Student, Industrial Engineering, University of Houston, E206 Engineering Bldg 2, Houston, TX, 77204, United States of America, gkaya@central.uh.edu, Gino Lim Intensity modulated proton therapy (IMPT) usually operates a constant relative biological effectiveness (RBE). In fact, RBE is not constant. RBE is described as a function of dose, linear energy transfer (LET) and tissue type in the structure of the linear-quadratic (LQ) model. We study the optimization of radiobiological effects (dose and rbe-weighted dose) in the context of LQ model by using two conjugate gradient algorithms. For results, we use data for head and neck cancer case. 3 - Robust Surgery Scheduling with Exception Analytics Yooneun Lee, The Pennsylvania State University, 236 Leonhard Building, University Park, PA, 16802, United States of America, yxl5250@psu.edu, Vittaldas Prabhu In this study, we address a surgery scheduling problem with uncertain surgery duration where surgical procedure takes place in multiple operating rooms. We present a robust surgery scheduling model and study its performance using exception analytics approach. We perform numerical experiments to compare performances of various models including simple heuristics, and find out that the results illustrate that the robust models with exception analytics works well across different instances. 4 - Intertemporal Decisions in Hospital Capacity Planning Jorge Vera, Professor, Universidad Catolica de Chile, Dept. Industrial and System Engineering, Vicuna Mackenna 4860, Santiago, Chile, jvera@ing.puc.cl, Ana Batista Correct planning of capacity in a hospital is crucial for high standards of service to patients. The problem is complex not only because of the different areas in a large hospital but also because of several uncertainties present in the system, like patient demand or length of stay. In this work we show how we could use an intertemporal hierarchical decisions modeling to address this problem. We present model alternatives as well as solution methods based on Stochastic Optimization 5 - Workload Balancing Problem in an Outpatient Center under Uncertanity Kamil Ciftci, PhD Candidate, Lehigh University, H.S. Mohler Laboratory, 200 West Packer Ave., Bethlehem, PA, 18015, United States of America, kac208@lehigh.edu Creating fair nurse workload in infusion center is a difficult task due to uncertainty in patient late cancelation and no-show while patient satisfaction is top priority for hospital. In this study, we propose two-stage stochastic program model to find best combination of nurse workload balancing schedule (NWBS) and patient waiting time (PWT) under different uncertainties. Computational results show that proposed methodology provides better NWBS and keeps average PWT under hospital goal. MC38 38-Room 415, Marriott Dynamic Programming and Control I Contributed Session Chair: Jefferson Huang, Stony Brook University, Dept. Applied Mathematics & Statistics, Stony Brook, NY, 11794-3600, United States of America, jefferson.huang@stonybrook.edu 1 - Dynamic Pricing Mapreduce Model Minghong Xu, Doctoral Student, University of Illinois at Chicago, 600 S. Morgan St., Chicago, IL, 60607, United States of America, summerinxu@gmail.com, Sid Bhattachary, Kunpeng Zhang Dynamic programming breaks the problem down into a collection of simpler subproblems and has the optimal substructure. But it suffers from “curse of dimentionality”. On the other hand, distributed implementation using MapReduce has been proved to be an efficient tool that solved a lot of large-scale problems. In this study, the Big Data era technics is used to solve Big State Space. A MapReduce Model is proposed for a Dynamic Pricing problem using E- Commerce data. 2 - Energy Storage Management in Microgrids: A Supplier’s Perspective Arnab Bhattacharya, University of Pittsburgh, 6236 Fifth Avenue, Apt. 102A, Pittsburgh, PA, 15232, United States of America, arb141@pitt.edu, Jeffrey Kharoufeh We consider a renewable energy supplier’s problem of optimally procuring, selling and storing energy when renewable supplies and real-time prices are uncertain. A finite-horizon MDP model is formulated and solved to maximize the supplier’s total expected (discounted) net profits, subject to storage capacity and transmission constraints.

217