Minimizing Delay at Closely Spaced Signalized Intersections Through Green Time Ratio Optimization: A Hybrid Approach With K-Means Clustering and Genetic Algorithms

Closely spaced intersections play a key role in traffic flow management. This study aims to model different traffic related parameters to minimize the delay of a closely spaced intersection by optimizing the green time ratio with the help of the genetic algorithm. The factors influencing optimization were selected as the distance between two adjacent intersections, cycle time, degree of saturation, green time ratio, volume, and the queue length-to-distance ratio, which is considered the Index parameter. The dataset was calibrated, validated, and used to simulate the analysis of traffic scenarios using SIDRA Intersection. To provide a clearer analysis, the k-means clustering algorithm was applied to divide the distances into three clusters. Among these clusters, the distance range between 110 and 160 meters is identified as the transition zone. The optimal green time ratio to minimize delay value for closely spaced intersection clusters was determined within a range of 0.58 to 0.69. To ensure a more comprehensive analysis, these values are used to examine their impact on delays. For this reason, the scenarios were restructured again with SIDRA using the newly optimized values to evaluate whether there is any reduction in the traffic-related parameters. The delay values and their temporal fluctuations showed significant improvements with this hybrid approach. The optimized green time ratios reduced delay, degree of saturation, and CO2 emissions by 8.95%, 8%, and 4.72% at the downstream intersection, and by 6.86%, 6.16%, and 7.09% at the upstream intersection, respectively.