Decision support system for road traffic management can be used for freight transport, people transport but also for site evacuation. We deal with two aspects of the decision support system in a same global architecture: one for traffic regulation to avoid jam and the other for road users to choose the shortest path in time between two points. These two aspects interact. The cartography is represented by a weighted digraph. The weights evolve according to the traffic and the graph is therefore dynamic. The regulation system is based on a neural network. The shortest path is based on an ant algorithm well suited for dynamic environments.