Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MB28

5 - Results from an Analytics Benchmarking Study Andrew Urbaczewski, University of Denver, Daniels College of Business 593, 2101 S. University Boulevard, Denver, CO, 80208, United States I will present the results from my Analytics Benchmarking Study. I have looked at dozens of Business Analytics/Data Analytics/Analytics/Data Science/etc. Programs and compared tuition, pre requisites, credit hours, and expected length to complete. n MB28 North Bldg 221A Yard and Terminal Operations Sponsored: Railway Applications Sponsored Session Chair: Tyler Dick, U. of Illinois at Urbana-Champaign, Urbana, IL, 61801, United States 1 - The Digital Transformation of Rail Yard Planning and Operations Jeremiah Dirnberger, GE Transportation, 7572 Old Kings Rd S, Jacksonville, FL, 32217, United States Rail yards are vital to overall network fluidity but have had limited technological investment relevant to other areas. GE Transportation is advancing a holistic solution to enhance reliability, reduce maintenance overhead, improve utilization of existing capacity, and increase margins and volume through these critical nodes. Information is analyzed at the edge and passed on for use by an integrated suite of inventory management and decision support tools. The work impact at downstream yards on and off each network are considered and decisions automatically made to improve productivity, reliability and flexibility, while enabling safer operations through effective use of automation. 2 - Improving the Flat Switching Process Roger Baugher, TrAnalytics, Johns creek, GA, United States, Daril Vilhena One of the most daunting challenges in many yards and terminals is developing and implementing efficient switching sequences - the process of taking cars from multiple tracks and building blocks most expeditiously while ensuring train makeup rules are enforced. The challenge is especially daunting in medium-sized to large flat yards. Railroads have spent heavily to implement yard management systems, but these systems typically provide little guidance on how to switch the cars. Excellent yardmasters can perform this mathematically-challenging task well, but their population is dwindling as experienced railroaders retire. The author will discuss the development of a set of new tools that provides new Tyler Dick, U. of Illinois at Urbana-Champaign, 1241 Newmark Lab MC-250, 205 N. Mathews Avenue, Urbana, IL, 61801, United States In designing a train plan for a railway network, one decision is to determine the number of blocks to be handled by each train and the number of blocks to be assembled at each classification yard while maintaining a certain level of service. This research uses Optym YardSYM to investigate the relationship between throughput volume, the total number of blocks assembled and the level of service at a hump classification yard. To supplement this initial analysis, the research also investigates the influence of number of departing trains, block size distribution and schedule and volume variability on yard performance. 4 - Car-Scheduling Based Hump Sequencing Roger Baugher, TrAnalytics LLC, Johns Creek, GA, United States, Chip Kraft In a car-scheduling driven approach to hump sequencing, the goal is not to make all connections, but to ensure that at least the most important ones are protected. Yard management uses the tool to decide whether to hold outbound train departures for a short time, or else drop the connection if the delay would be too long. capabilities and improves yard performance in both time and cost. 3 - Traffic Complexity and the Performance of Railway Classification Yards

n MB29 North Bldg 221B Information and Preferences in Traffic Flows Sponsored: TSL/Urban Transportation Sponsored Session Chair: Laiyun Wu, SUNY-Buffalo, 326 Bell Hall, Buffalo, NY, 14226, United States 1 - Modeling Spatiotemporal Information Flow Propagation in a Vehicle-to-vehicle- Communication System Considering Communication Delays Yangjiao Chen, Purdue University, West Lafayette, IN, United States This study develops an analytical Markov model to characterize the spatiotemporal propagation of information under vehicle-to-vehicle (V2V) communications while factoring traffic dynamics and communication delays due to communication failure and communication frequency. A closed-form solution of the expected information propagation speed is derived under different densities of equipped vehicles. Numerical experiments demonstrate the effectiveness of the proposed model in various traffic conditions. 2 - Evaluating the Cognitive Effects of Real-time Travel Information using Psychophysiological Analysis and their Implications for Driver Decision-making Shubham Agrawal, Purdue University, West Lafayette, IN, 47906, United States This study conducts interactive driving simulator-based experiments to evaluate the impacts of driver cognitive state (for example, mental workload and engagement level) on the driver route choice decision-making process under real- time travel information provision. The driver cognitive state is estimated by analyzing the physiological data collected using electroencephalogram (EEG), electrocardiogram (ECG) and wearable eye-tracking glasses. The systematic differences in driver cognitive state are analyzed based on the characteristics of the disseminated real-time information and heterogeneity in individual characteristics. 3 - Expectations of the Driver’s Role when Using an Automated Driving System Dustin Souders, Purdue University, West Lafayette, IN, United States This study investigates the effects of introductory materials on participants’ interaction with an automated driving system (ADS; SAE level 3) in a simulated environment. Young and old participants are engaged in a secondary task while monitoring an ADS, and vigilance patterns (eye-tracking, EEG), take-over performance, trust and acceptance attitudes will be assessed. Results will inform how licensing agencies and OEMs should train drivers when using level 3 automation during this transitionary period in road vehicle automation to ensure proper expectations and encourage safety. 4 - Cooperative Adaptive Cruise Control for Connected Autonomous Vehicles by Factoring Communication-Related Constraints Chaojie Wang, Purdue University, West Lafayette, IN, United States We propose cooperative adaptive cruise control (CACC) strategies for connected autonomous vehicles (CAVs) to enhance platoon performance by temporarily switching off the V2V communication functionality for some CAVs in the platoon. An optimization model is established to determine the optimal information flow topology that can maximize platoon performance based on one or more objectives (string stability, smoothness and convergence rate of the platoon control strategy) under the communication-related constraints. The effectiveness and efficiency of the proposed CACC strategies for CAV platoons will be illustrated using numerical simulation. 5 - Inferring Origin-Destination and User Preference in Multi-modal Travel Environment by Using Automated Fare Collection Data Laiyun Wu, SUNY-Buffalo, 326 Bell Hall, Buffalo, NY, 14226, United States The Origin-Destination (OD) demand data availability and quality are critical for the effective and efficient operation and management of a transit system. Understanding of transit user preferences, at the same time, is also important for planning and assessing transit systems. In this paper, we develop and apply an inference framework for distilling multi-modal routing preferences for transit system users and their true OD through a probabilistic learning method, based on a real-world Automated Fare Collection data set.

158