Determining that strategically placing restrictions on left-hand turns in high traffic areas would reduce traffic bottlenecks, researchers from Penn State University have developed a method for identifying where those restrictions would have the greatest impact.

According to the research team, left-hand turns are associated with increased traffic and accidents. As such, the team relied on a combination of two heuristic algorithms — a population-based incremental learning (PBIL) algorithm and a Bayesian optimization algorithm — to dictate where left-hand turns could be eliminated in high-traffic areas to achieve more efficient traffic flow.

The new hybrid approach developed by the Penn State researchers combines the PBIL algorithm, with its randomly sampled potential configurations for recognizing the patterns of high-performing options, with the Bayesian optimization algorithm, which analyzed the patterns to identify how restrictions impacted traffic at adjacent intersections.

"Instead of starting the Bayesian optimization with a random guess, we fed it with the best guesses from the PBIL," researchers explained. "The first method creates the starting point, and the second refines it."

The hybrid method was tested via a simulated, square network under assorted scenarios and the team determined that all three methods — PBIL, Bayesian optimization and hybrid — identified configurations that resulted in more efficient traffic patterns, more so than a design without restrictions.

The researchers also found that for simulations with more realistic settings, the hybrid method proved to be the most effective, while the most efficient configurations banned left turns in the middle of the city while allowing them on the periphery of the city.

The study appears in the Transportation Research Record: Journal of the Transportation Research Board.

To contact the author of this article, email