The IE Research Datalab continues its seminar series with a presentation by Miguel Vaquero on optimizing traffic flows through data-driven techniques. Professor Vaquero combines advanced simulations with mathematical optimization algorithms, resulting in a system that manages urban traffic more efficiently.
Miguel Vaquero began by illustrating how his simulations can be used to learn the dynamics of traffic flow. He then proved that managing this flow is a non-convex optimization problem. Non-convex problems have more than one optimal solution, but only one of them is the global optimal! This property makes non-convex problems a hard challenge to solve. Miguel Vaquero proposes a modification of primal-dual dynamics to tackle these kinds of optimization problems.
A key aspect of Miguel Vaquero’s presentation was the emphasis on the mathematical guarantees of his algorithm, focusing on tracking and robustness. These guarantees are critical, as they ensure the algorithm’s performance remains stable and reliable under a wide range of conditions, making it a viable solution for actual traffic management systems.
To wrap up the seminar, Miguel Vaquero showed a comparative analysis between traditional traffic management methods and his optimized approach.
Traditional traffic management: https://www.dropbox.com/s/fr1w7yt5aii680a/mult.avi?e=1&dl=0
Professor Vaquero’s optimization: https://www.dropbox.com/s/885jxze75mj5u0h/mult_opt.avi?e=1&dl=0
The number on the top left corner shows the number of cars leaving the intersection, which is the objective function. This data shows a significant increase in the number of cars able to leave an intersection when using professor Vaquero’s optimization technique.