Traffic Signal Control by Using Traffic Congestion Prediction Based on Pheromone Model

Traffic congestion has been a long-held social problem because of the increasing number of vehicles. In this paper, we apply a pheromone model to a traffic signal control in order to alleviate the traffic congestion. The pheromone model is a tool to communicate among insects (e.g., ants) for their crowd action. In our target problem, the pheromone is strewed by vehicles in a trail across the road instead of insects, and the amount of pheromone corresponds to the degree of traffic congestion. Moreover, the factors of traffic signals (i.e., cycle, split, and offset) are controlled by the pheromone to reduce the idling time in front of traffic signals. We performed multi-agent simulations to compare our traffic control method with other control methods. The results show that our proposal method has good effects to the traffic congestion problem.