A Unified Software Framework to Enable Solution of Traffic Assignment Problems at Extreme Scale

We describe a modular software framework for solving user equilibrium traffic assignment problems. The design is based on the formulation of the problem as a variational inequality. Unlike most existing traffic assignment software which focus on specific traffic models, our framework accommodates various traffic models, but also enables using parallel computation in high performance computing environments to speed up large-scale equilibrium calculations. We compare the solutions obtained under several models: static, Merchant-Nemhauser, `CTM with instantaneous travel time', and `CTM with actual travel time', and demonstrate some important differences. We also present results for the Los Angeles Basin network, which demonstrate a 25x speed up for the parallelized static equilibrium calculation compared to the sequential computation.