Traffic Signal Control Based on Game Theoretical Framework

This paper presents a model of intelligent traffic signal control using a game theoretical framework based on decision-making operations. This model aims at finding an optimal solution to reduce traffic congestion in the urban transportation system. Thus, the strategies are generated to maximize the number of vehicles passing through an intersection within a cycle time. Urban traffic is mostly controlled by signal lights, so optimum traffic management can be achieved by making strategies of the light signals according to the distribution of vehicles in the intersection. Incoming links (i.e. sections) in the intersection are regarded as players and the status of signal light (green or red) can be considered as the decisions made by players. In this way, the method is like a cooperative game combining all the decisions and evaluating a scalar cost function. Furthermore, the collision among vehicles must be taken into account while many vehicles are entering the intersection from different incoming links. At last, the effectiveness of one intersection is compared between the optimal strategies produced from a game theoretical framework and constant strategies implemented in real traffic models.