2016 INFORMS Annual Meeting Program

SD70

INFORMS Nashville – 2016

5 - Prioritizing Hepatitis C Treatment in U.S. Prisons Turgay Ayer, Georgia Tech, Atlanta, GA, Boston, MA, ayer@isye.gatech.edu , Anthony Bonifonte, Can Zhang, Anne Spaulding, Jagpreet Chhatwal High prevalence of HCV in prisons offers a unique opportunity to control the HCV epidemic. Newest HCV treatments drugs are effective but providing treatment is outrageously expensive. We propose a restless bandit modeling framework to support hepatitis C treatment prioritization decisions in U.S. prisons. From the interpretation of this closed-form expression, we anticipate the performance of Whittle’s index would degrade as the treatment increases. Using a detailed agent- based simulation model, we show our proposed policy can significantly improve overall health outcomes compared with the current practice. Our results shed light on issues in hepatitis C prioritization: 1) considering remaining sentence length and injection drug use (IDU) status and liver health state in prioritization decisions can lead to a performance improvement; 2) when linkage-to-care rate outside prison is small while treatment capacity in prison system is relatively large, patients with shorter remaining sentence lengths should be prioritized; and 3) for patients with advanced liver disease, IDUs should not be prioritized unless their reinfection is very-well controlled. SD70 Acoustic- Omni Transportation, General Contributed Session 1 - Solving The Privately Owned Automated Vehicles Assignment Problem Theresia van Essen, Delft University of Technology, Mekelweg 4, Delft, 2628 CD, Netherlands, j.t.vanessen@tudelft.nl, Gonçalo Correia We propose a new model to study how replacing privately owned non-automated vehicles with shared automated ones affects travel time, congestion and parking demand in an urban area. As automated vehicles will reduce the value of travel time, it is expected that travel time will increase. In addition, congestion is on the one hand expected to increase because of the empty trips, and on the other hand expected to decrease because of a reduction in the number of vehicles on the road. Parking demand is expected to decrease as the utilization of the vehicles will increase. The model is applied to a case study based on the city of Delft, the Netherlands. 2 - Developing Interrelated Airport Facilities under Uncertainty: A Network Flow Formulation Yanshuo Sun, University of Maryland, 1173 Glenn Martin Hall, College Park, MD, 20742, United States, yssun@umd.edu, Paul Schonfeld Interactions between user flows and facilities are quite complex in an airport system. Thus, capacity expansion decisions for these facilities are largely interrelated. A network flow formulation is proposed for coordinating such development decisions so that a balanced capacity configuration is likely to be obtained. The nonlinear congestion effect, which is common in most airport facilities, is considered and uncertainties in demand and aircraft mix are also included. The stochastic mixed integer nonlinear program is reduced to a deterministic mixed integer program and thus solved. 3 - From Trend Spotting To Trend Setting: Modeling The Impact Of Major Technological And Infrastructural Changes On Travel Demand Feras El Zarwi, PhD Candidate, University of California at Berkeley, 2100 Channing Way, Apt 415, Berkeley, CA, 94704, United States, feraselzarwi@gmail.com Transformative mobility will revolutionize travel and activity behavior but we should be cautious with how the future is going to play out. This research project proposes a methodological framework tailored to address impacts of technological innovation to understand and predict long-range trends in travel behavior. We integrate hidden markov and discrete choice models to predict long-range trends in travel behavior as a result of adopting new services. The model is estimated on a longitudinal travel diary dataset from Santiago, Chile. The proposed quantitative methods are critical in assessing how policies/strategies can influence trends of travel behavior to guide transformative mobility.

4 - Using Regression Tree Models To Improve Freeway Incident Duration Prediction A Comprehensive Case Study In Maryland Region Xuechi Zhang, Graduate Research Assistant, University of Maryland, 0147C Engineering Lab Building, College Park, MD, 20740, United States, zhangxc90@gmail.com, Ali Haghani, Yeming Hao Timely and accurate prediction of freeway incident duration is not only useful for providing travelers with re-routing strategies, but can also reduce their in-vehicle anxiety. This research proposed several regression tree based models to improve the incident duration prediction accuracy by fusing heterogeneous information, i.e. incident information, weather and traffic conditions. A comprehensive case study with real-world data in Maryland Region was conducted to evaluate and demonstrate the proposed models. Further, practical implications from the case study were given. SD71 Electric- Omni Transportation, Rail Contributed Session Chair: Emmanuel Martey, University of Delaware, 302 DuPont Hall, Newark, DE, 19716, United States, enmartey@udel.edu 1 - Optimization Techniques For Railways Srinivasa Prasanna, Professor, IIIT-Bangalore, 26/C, Hosur Road, Electronics City, Opposite Infosys Technologies, Bangalore, 560100, India, gnsprasanna@iiitb.ac.in” We present optimization techniques used in portions of the Indian Railway System, the largest in the world. We present techniques used for timetabling and investment planning, under large scale demand uncertainty. Many of these problems are at the scale of grand computational challenges (with 10’000 of trains and 1000’s of control points), and the talk will present a few pieces of how portions of this problem can be simplified and made amenable to optimization techniques (convex/non-convex). Exemplary results will be discussed. 2 - Railway Capacity Analysis And Cyclic, Combined Train Timetabling And Platforming For A Single Track, Bidirectional Railway Line Matthew Petering, University of Wisconsin-Milwaukee, Industrial and Manufacturing Engineering Dept, Ems E367, Milwaukee, WI, 53201, United States, mattpete@uwm.edu, Mojtaba Heydar We present the literature’s first mixed integer linear program for cyclic train timetabling and platforming on a single track, bi-directional railway line. There are T train types and one train of each type is dispatched per cycle. The decisions to be made include the sequencing of the train types on the main line and the assignment of train types to station platforms. Two conflicting objectives— minimizing cycle length and minimizing total train journey time—are considered. 3 - A MIP Model For High-speed Train Platforming Problem With Route Conflicts Constraints Gongyuan Lu, Assistant Professor, Southwest Jiaotong University, 111#, 1st Section, Northern 2nd Ring Road, School of Transportation and Logistics, Chengdu, China, lugongyuan@swjtu.cn, Guangyuan Zhang, Yuan Wang The biggest challenge in solving high-speed train platforming problem (HTPP) is to route trains without conflicts. Especially in a large multi-yard railway station, the conflicts between routes and platforms can easily make the scale increase dramatically. A MIP model aiming at minimize train delay is formulated to generate flexible train schedule without violating route conflicts. This research has been applied in largest high-speed train station in Asia. 4 - Predicting Cascading Effects Of Local Disruptions In A Large Scale Rail Network Patrick Briest, McKinsey & Co, Kennedydamm 24, Dusseldorf, 40027, Germany, patrick_briest@mckinsey.com, Sebastian Albert, Robin Blöhm, Florian Brummer, Christian Gruß, Eike-Dennis Rausch We propose a stochastic simulation model to determine network-wide effects of locally induced disruptions in a large scale rail system. We extend the model previously described by Berger et al. to include (A) dynamic diversion routing to mimic how traffic controllers will try to route trains around disrupted parts of the network and (B) track capacities and a load-dependent delay component. We present computational results based on detailed delay distributions extracted from multiple years of Deutsche Bahn’s operations data and simulation runs using both current and historic schedules of regional, long-distance and cargo traffic in the Deutsche Bahn network.

118