Abstract

On the use of high-resolution urban traffic simulation models for large-scale network optimization

Carolina Osorio
Massachusetts Institute of Technology

In this talk we present a family of optimization methods that enable the use of high-resolution stochastic traffic simulators for optimization. These methods are known as simulation-based optimization algorithms. We present algorithms for large-scale dynamic problems, as well as ongoing work for real-time problems. We use these algorithms to address urban traffic management problems, and present the results of collaborations with various transportation agencies. We discuss recent results on the design of urban traffic management algorithms that account for the presence of (semi-)autonomous vehicles.