Simheuristic Framework for Optimizing Urban Mobility at Signalized Roundabouts

Abstract

Managing high traffic volumes and traffic congestion at signalized intersections remains a critical urban challenge. Appropriate traffic signal timing (TST) and phase sequencing are essential for ensuring smooth traffic flow. This study presents a microscopic simulation-based heuristic optimization (Simheuristic) framework using the Genetic Algorithm (GA) for optimizing the TST of Four-Legged Two-stops Signalized Roundabouts (FLTSR). The framework is tested using the actual traffic flow through a microscopic simulation model developed in Simulation for Urban Mobility (SUMO). Within this framework, the integrated GA searches for the green TSTs to minimize vehicular queue lengths, while SUMO is used to evaluate those timings. Additionally, four different phase sequence settings are evaluated to find the efficient configuration. The proposed approach is benchmarked against Webster's method and the existing TST plan. In the best-case scenario, the proposed framework improves vehicular flow by mitigating the average time loss, average waiting time, and the average number of vehicles in a queue at the FLTSR up to 35.83 %, 51.91 %, and 50.97 %, respectively, compared to the current setting. (Received in November 2025, accepted in January 2026. This paper was with the authors 1 month for 1 revision.)