Evaluation of heuristic algorithms for traffic signal optimization along a corridor

Traffic signals are one of the most important parts of road networks. They play a significant role in decreasing congestion and providing better service for the society in urban areas. The aim of this thesis is to determine the best traffic signal timing plan using heuristic algorithms for a corridor consisting of several intersections with a microscopic traffic simulator based on a car following theory, namely Intelligent Driver Model. The measure of effectiveness used in this study is the duration of stopped vehicles. All available signal timing plans for the corridor are evaluated and the results illustrate that determining the best traffic timing plan can effectively reduce the congestion in corridors. Based on the evaluation of the performances of heuristic algorithms (Simulated Annealing, Hill Climbing, Genetic Algorithm), Genetic Algorithm provides the global optima for all scenarios for this specific corridor. In this study, it is also determined that Hill Climbing Algorithm works 28 times faster than Genetic Algorithm, but it fails to reach global optima in the predetermined time duration. Genetic Algorithm's search process takes much more time than the Hill Climbing and Simulated Annealing; however, its performance for determining global optima in the given number of generations makes it the best option in this study.