Optimization of Multi-Junction Traffic Light Control Using the Classic Genetic Algorithm

As traffic congestion is a known problem globally, including Malaysia, it is vital to investigate the feasibility of optimizing the traffic light controls across multiple linked junctions. This paper explores the implementation of a genetic algorithm (GA) to optimize the traffic light signal durations of multi-junction traffic light control systems to reduce traffic congestion, using the junction network in Bandar Sunway as a case study. This paper aims to formulate the problem of traffic light signal durations across multiple junctions to be solved with the classic GA and evaluate the performance of the optimized timing with the existing traffic light timing. The GA is designed to identify the optimal signal durations for light, medium, heavy, and very heavy traffic, respectively, at Bandar Sunway. Simulation is performed to evaluate the improvements in the GA durations over the existing durations. This paper proposes a simplified calculation of the vehicle travel time as the objective function to evaluate the fitness of the individuals in GA. The case study presented in this paper focuses on the linked junctions with the main traffic flow in Bandar Sunway. The performance of GA is compared to the fixed-time signaled traffic light, of which the timing is defined based on the actual traffic light timing implemented at the junctions in the real world. The results show a significant improvement of a 10 to 25% reduction in average waiting time, and a 10 to 30% reduction in maximum waiting time upon using the GA. Thus, the classic GA is proven to provide better sets of fixed traffic light timing at different traffic conditions which improves the traffic flow.