2016 INFORMS Annual Meeting Program

WB89

INFORMS Nashville – 2016

2 - A Multi-level Approach To Network Attack Graph Interdiction David Joseph Myers, Research Engineer, Air Force Research Laboratory, AFRL/RISB, Attn: David Myers, Rome, NY, 13441-4505, United States, david.djm.myers@gmail.com Attackers have a distinct advantage in the cyber domain, having unlimited time to perform reconnaissance on an enterprise system. A key component of the defender’s ability to protect their system is a situational awareness about the system attack graph. The defender’s goal is to minimize the potential damage through the exploitation of vulnerabilities. This presentation will explore the development of these attack graphs, and then the consideration of three approaches to interdicting this graph. 3 - Evaluating Basing Options For Optimizing Accessibility For Global Response Force Jeremy Eckhause, Operations Researcher, RAND Corporation, 1200 S. Hayes St., Arlington, VA, 22202, United States, eckhause@rand.org, Katharina Ley Best, Christopher Pernin, Michael Schwille, Katherine Pfrommer For a global response force to achieve its mandate, rapid access to almost any point on the globe is essential. Since the long-term presence of the US is difficult to predict, using of a set of intermediate bases may be required for establishing fast and sustainable access to large numbers of contingency locations. We present an approach and results for identifying a robust set of intermediate bases for ensuring global access and a methodology for identifying new bases as infrastructure requirements change. 4 - Predicting Future World Conflict Using Factor Sample Paths Darryl K Ahner, Air Force Institute of Technology, 135 Eastwick Court, Dayton, OH, 45440-3647, United States, darryl.ahner@afit.edu, Nicholas Jerred Shallcross The prediction and forecasting of nation conflict is of vital importance. This paper discusses the formulation and construction of a suite of region-specific conditional logistic regression models that predict nation-state transitions into and out of violent conflict. This approach allows for the accurate modeling of complex regional environments with parsimonious and operationally interpretable models. The conditional logistic regression models proposed in this study achieve conflict transition prediction accuracies of 84.67% for 182 of the world’s nations. Several predictor variable paths are explored and their effect on probability of nations being in a state of conflict are analyzed. 5 - Assignment Of UAVs To Search And Communication Roles Michael Atkinson, Naval Postgraduate School, 1411 Cunningham Road, Building 302, Monterey, CA, 93943-5219, United States, mpatkins@nps.edu, Ezra Akin, Kevin D Glazebrook Once a search UAV detects a target, the UAV must transmit the target’s position to a shooter who will fire on the target. If the shooter is far away from the search area in a communication-degraded scenario, we may need several communication UAVs in intermediate positions to relay the message from the search UAV to the shooter. We examine the assignment of UAVs to tasks in real- time. On one hand we want as many UAVs searching so they can quickly detect as many targets as possible. On the other hand we need a sufficient number of communication UAVs to ensure a robust communication network. We formulate an MDP that provides optimal solutions for small problems and develop heuristics to use on larger problems. WB89 Broadway C-Omni Maritime Transportation Sponsored: TSL, Intelligent Transportation Systems (ITS) Sponsored Session Chair: Harilaos Psaraftis, Technical University of Denmark, Department of Transport, Lyngby, 2800, Denmark, hnpsar@dtu.dk Co-Chair: Dario Pacino, Technical University of Denmark, Denmark, darpa@dtu.dk 1 - The Ship Loading Problem With Straddle Carrier Assignment And Scheduling Dario Pacino, Technical University of Denmark, Ostrigsgade 28, Copenhagen, Denmark, darpa@dtu.dk The maritime shipping sector is under pressure to provide reliable and cheap services. Operations research techniques have caught the interest of the industry as can be seen from the increasing number of publications in e.g. liner shipping network design and terminal optimization. In this paper we proposed the Ship Loading Problem, a novel collaboration approach to integrate shipping line and container terminal cost optimizations. We present a novel mathematical formulation and a heuristic approach which demonstrates the benefits of this collaboration.

2 - Containership Deployment On A Liner Service Shuaian Wang, Hong Kong Polytechnic University, Department of Logistics and Maritime Studies, Hong Kong, China, hans.wang@polyu.edu.hk This study proposes an important ship voyage management problem (SVMP) that aims to minimize the bunker fuel consumption of a containership. To address the SVMP, we first develop a tailored method to build two robust artificial neural network (ANN) models using ship voyage report data to quantify the synergetic influence of sailing speed, displacement, trim, and weather/sea conditions on ship fuel efficiency. We proceed to put forward three viable solution countermeasures for the SVMP by means of dynamic programming and simulation-based optimization techniques. 3 - A Metaheuristic For a Multi-product Maritime Inventory Routing Problem We consider a multi-product maritime inventory routing problem where an actor is responsible for both the inventory management of the various products at the ports and the ships’ routing and scheduling. In addition, we take the allocation of products to undedicated compartments into account. A mixed integer programming model is formulated, and it can be solved to optimality for small instances only. A matheuristic, exploiting the two sub sets of constraints related to the routing and inventory management, is developed. The computational study shows promising results for the matheuristic. WB90 Broadway D-Omni Health Care, Modeling XIII Contributed Session Chair: Masoud Kamalahmadi, Doctoral Student, Indiana University, 1309 E Tenth St, Bloomington, IN, 47405, United States, maskamal@iu.edu 1 - Inventory Policies For Platelet Management At Hospitals Under Demand Uncertainty Suchithra Rajendran, PhD Candidate and Research Assistant, The Pennsylvania State University, 310 Leonhard Building, University Park, PA, 16802, United States, sur205@psu.edu, Arunachalam Ravindran Demand uncertainty at hospitals leads to a significant wastage of blood platelets. Hence, a stochastic mixed integer linear programming (SMILP) model is developed with the objective of minimizing platelet wastage and shortage. Due to the complexity of the SMILP, five different heuristic approaches are developed using historical data such as mean and variance of platelet demand. Real-life data from a Regional Medical Center is used to evaluate the different methods. In addition, sensitivity analysis is performed to evaluate the robustness of the proposed heuristics. The results indicate that the heuristic approaches on average, provide a solution within a gap of 10% from the optimal solution. 2 - An Optimization Framework To Improve Patient Safety In Radiation Therapy Pegah Pooya, PhD Candidate, North Carolina State University, 304 Ravenstone drive, Raleigh, NC, 27518, United States, ppooya@ncsu.edu, Osman Ozaltin, Julie Ivy, Lukasz Mazur, Lawrence Marks, Katharin Deschesne, Prithima Mosaly, Gregg Tracton The use of safety barriers (SB) in radiation therapy (RT) is a widely recognized method for detecting potential human and non-human errors before they reach patients. We develop an optimization framework to determine the reliable design of SBs to improve patient safety considering SB implementation costs. 3 - A Theoretical Agent-based Framework To Evaluate The Anticompetitive Implications Of Accountable Care Organizations Abdullah Alibrahim, PhD Candidate, University of Southern California, 344 Hauser Blvd Apt 219, Los Angeles, CA, 90036, United States, alibrahi@usc.edu, Shinyi Wu In the wake of Affordable Care Act, two initiatives have seemingly counteracting effects. Market share shifts due to coordinating healthcare provision (Accountable Care Organizations -ACOs) might negate the effects of concentrating purchasing for care and coverage (Health Exchanges). This study justifies characterizing healthcare markets as a complex adaptive system and outlines a framework to assess competitive implications of ACOs in private healthcare markets. The theoretical, structural, behavioral, and iterative relationships of the system are outlined. An agent-based simulation model will then be used to assess competitive effects of ACOs to inform antitrust policies. Marielle Christiansen, Norwegian University of Science & Technology, Industrial Economics & Technology Mgmt, Trondheim, Norway, Marielle.Christiansen@iot.ntnu.no

426