A high-resolution (HR) data system for a signalized intersection collects the location (lane), speed, and turn movement of every vehicle as it enters the intersection, together with the signal phase. The data can be used to estimate several performance metrics of the intersection. Historical data can be used to understand trends and to design optimal timing plans. Real time data can be used to make predictions several hours ahead and to change timing plans accordingly. The talk will describe an HR system and illustrate the uses of its data.

Autonomous and connected vehicles are likely to become an integral part of the traffic stream within the next few years. Their presence is expected to greatly modify travel demands and habits, traffic flow characteristics, traffic safety and related external influences, such as energy consumption, emissions and pollution. Already a significant body of research exists on the technical requirements, sensing and communications systems and the algorithms that could control a single autonomous and connected vehicle or coordinate movement of these vehicles when they are interacting with each other.

Evaluations of the implications of the presence of these vehicles in the traffic stream on traffic flow as a whole and on externalities are needed. This suggests a need to enhance current traffic models, in particular microscopic ones, to represent autonomous and connected vehicles, the algorithms that control and coordinate between them, their interactions with non-autonomous vehicles and the behaviours of human drivers in the presence of these vehicles. This talk would focus on defining requirements and specific traffic flow modelling needs that account of the presence of autonomous and connected vehicles, their specific physical attributes, modes of operation and characteristics of scenarios in which they will operate.

Abstract: We address a key problem in B2B and B2C supply chain logistics, particularly important in the volumes of movements in emerging economies. Both B2B and B2C transactions require pickups and deliveries of shipments between businesses (B2B) and to the end users (B2C). A core problem in these settings in the Vehicle Routing Problem with Time-Windows (VRPTW), studied extensively for the last five decades. To address issues of scalability for very large instances that emerging economies demand, we present a novel Deterministic Annealing (DA)-based approach to solving the VRPTW. Our approach models aspects of routing and scheduling, as well as additional constraints of heterogeneous vehicles and shipments. This is the first time, to our knowledge, that a DA approach, inspired from information theory, has been used for problems in the class of the VRPTW. We describe how our DA approach gives rise to an effective heuristic for the VRPTW, which is designed to not get trapped in local minima, and demonstrates less sensitivity to initial solutions. The algorithm trades off routing and scheduling in an n-dimensional space using a tunable parameter that allows us to generate qualitatively good solutions. Businesses can tune these parameters to get qualitatively good solutions that differ in specific properties and allow the manager the freedom to choose tailored solutions. Simulation results on benchmark instances demonstrate the scalability of our approach to very large instances, outperforming current methods in terms of speed, and also in quality for a majority of instances. This is joint work with Mayank Baranwal, Srinivasa Salapaka and Carolyn Beck.

Bio: Lavanya Marla is an assistant professor in the department of Industrial and Systems Engineering at the University of Illinois at Urbana-Champaign. Prior to this, she was a Systems Scientist with the iLab’s Mobility Analytics group in the Heinz College at Carnegie Mellon University; and she holds a PhD in Transportation Systems from the Massachusetts Institute of Technology. She develops models and algorithms for large-scale transportation and logistics systems, for robust decision-making under uncertainty, real-time adaptability, using tools from optimization, statistics, simulation, simulation-optimization and other computational techniques. She has applied her work to aviation, logistics, emergency management and shared transportation systems. Her recognitions include a US Department of Transportation award for a University Transportation Center at CMU, and the US-India Educational Foundation's 21st Century Knowledge Initiative grant. Her experience has involved collaborations with IBM Research, UPS, NASA, various airlines, and consulting companies.