Ketan Savla, of University of Southern California, presented Horizontal Traffic Queues at the ITS Transportation Seminar March 15, 2019 at 4 p.m. in 290 Hearst Memorial Mining Building.
The service paradigms typically used in queuing frameworks for transportation systems do not have sufficient resolution to model emerging automation technologies at the microscopic scale. Such inadequacies can translate into erroneous performance estimates at the system scale, e.g., in terms of capacity, travel time, or waiting time. We advocate a state-dependent queuing paradigm, and present our work in this context for canonical transportation settings. Specifically, we generalize existing notions of spatial, vacation, and processor sharing queues to enable accurate performance evaluation for dynamic vehicle routing, signalized traffic intersections, and freeway traffic, respectively. A recurrent theme is to identify limiting scenarios that lend themselves to analytical or computational tractability, and compare against existing methodologies in practical regimes.