Optimizing PCU Values at Roundabouts Using PSO-Based Machine Learning Algorithm

Machine Learning tools are extremely adaptable in traffic engineering, providing innovative solutions to complex problems. This study investigates into utilising of Particle Swarm Optimisation (PSO) to create a predictive machine learning model for calculating Passenger Car Unit (PCU) values at roundabouts under heterogeneous traffic conditions. Applying PSO, the model attains a logically consistent set of PCU values, showcasing higher accuracy in comparison to real-world counterparts. The optimization process was employed using convergence analysis to ensure the model's parameters converge, resulting in a reliable and robust outcome. The model exhibited outstanding performance with an R-square of 0.986 at the 290<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">th</sup> iteration out of 1000, achieving a mean square error (MSE) of approximately 7%. The study demonstrates how PSO can optimize PCU values for roundabouts in various traffic scenarios, improving our comprehension of traffic flow dynamics at these important intersections.