Reversible discrete event formulation and optimistic parallel execution of vehicular traffic models

Abstract

Vehicular traffic simulations are useful in applications such as emergency planning and traffic management, for rapid response and resilience. Here, a parallel traffic simulation approach is presented that reduces the time for simulating emergency vehicular traffic scenarios. We use a reverse computation-based optimistic execution approach to parallel execution of microscopic, vehicular-level models of traffic. The unique aspects of this effort are: exploration of optimistic simulation of vehicular traffic; addressing the related reverse computation challenges; achieving absolute, as opposed to self-relative, speedup. The design, development and performance study of the parallel simulation system is presented, demonstrating excellent sequential and parallel performance. A speed up of nearly 20 on 32 processors is observed on a vehicular network of 65,000 intersections and 13 million vehicles.