Designing a queue events system for road traffic simulator

The urban traffic is frequently perturbed by congestions, followed by usual delays, accidents and road closures that cause supplementary delays. It is important to use intelligent systems for traffic simulation, routing strategies and scheduling techniques in order to achieve an optimal usage of resources. The multi-agents technology gives the advantage of implementing a system that can learn from his experience. Scheduling of message queues has strong similarities with algorithms used for giving different service rates to flows within network routers. In this paper a message scheduler that uses time-based credit-scheduling is proposed, in which resources are partitioned among messages solvers by allocating credits. Various aspects of the overall system are investigated in particular in relation to throughput, delay, scalability with processors and performance isolation.