This research investigates the application of different deep reinforcement learning methods for optimizing traffic light control in multi-modal urban traffic environments using the SUMO traffic simulator. Urban traffic congestion, with its significant economic, environmental, and social impacts, necessitates more sophisticated control strategies that can adapt to varying traffic conditions. Traditional traffic control systems, like fixed-time and adaptive methods, are often insufficient in handling the complexity of multi-modal traffic, which includes various traffic modes such as passenger cars and buses. Deep reinforcement learning, with its ability to dynamically optimize traffic light control without requiring prior knowledge of traffic patterns, is a promising method to improve traffic efficiency and achieve transit priority in multi-modal traffic. The research aims to address the limitations of the existing relevant research by employing several deep reinforcement learning algorithms, particularly multi-agent deep reinforcement learning methods, to coordinate multiple traffic lights in SUMO simulation. Research experiments are conducted in three different cases, which are set in road networks of different sizes respectively, and fixed traffic light control and max-pressure traffic light control are implemented for comparison. The applied deep reinforcement methods are evaluated in terms of training process and model evaluation. And the research results demonstrate that deep reinforcement learning methods, especially multi-agent deep reinforcement learning methods, can significantly enhance traffic flow efficiency and achieve transit priority in complex urban settings in multi modal simulation.

Urban areas with many daily commuters often experience a reduction in traffic flow during rush hours. In some urban areas where this reduction in traffic flow occurs, there are popular destinations for leisure that have multiple large parking facilities. Vehicles searching for vacant parking spots drive at a lower vehicle speed, reducing the traffic flow even further. Traffic flow can be improved by using a traffic signal control strategy and a smart parking solution to increase the speed of vehicles searching for vacant parking spots. This thesis combines the use of Model Predictive Control (MPC) in traffic signal control, with a Parking Resource Allocation Model (PRAM) to further improve the traffic flow on the traffic network. The PRAM considers the travel time of all travel routes to the parking areas, as well as the distribution of vehicles toward the parking areas.
The performance of the combined MPC and PRAM control strategy is compared using two case studies. In both case studies, the traffic network is a simplified representation of the mall of the Netherlands, located near Leidschendam. The first case study simulates traffic based on the morning rush hours, and the second case study simulates traffic based on the evening rush hours. In both case studies, multiple traffic demands are simulated, based on historical traffic data provided by Rijkswaterstaat. No historical data is provided by the mall of the Netherlands so fictive data is used. The performance of an MPC control strategy, an MPC control strategy with the first PRAM, and an MPC control strategy with the second PRAM is compared with a fixed-time control strategy for these traffic demands. The results show that the MPC control strategy reduces the total time spent and vehicle time loss of all the vehicles for the morning rush hours, but not for the evening rush hours. For one traffic demand, the added PRAMs further reduce the total time spent and vehicle time loss of all the vehicles, in both case studies. For the other traffic demands, the added PRAMs increased the total time spent and vehicle time loss. There is no significant difference between the two PRAMs in terms of the total time spent and vehicle time loss. Furthermore, the distribution of vehicles to the parking areas is more evenly for the second PRAM. Since the parking demand is fictive, future research is necessary with accurate parking demand to ensure that the combined MPC and PRAM further improves the traffic flow on the traffic network of the mall of the Netherlands. Moreover, future research is needed to more accurately predict the travel time of travel routes.