Impact of multi-criteria routing on dynamic traffic management

Traffic simulation is conventionally based on travel time as input to the cost function. We include vehicle emission into cost function and try to investigate its impact on traffic management as well as network performance. As the emission is the product of vehicle movement, we use a bi-level approach to model both the cost (with travel time and emission) and the flow. Experiments are first carried out at a Braess network to illustrate principal ideas and DTA simulations are then performed on an urban network to find out traffic performance at a real network. Our main research question consists of finding out what impact of multi-criteria cost function has on network traffic performance and thus traffic management, and whether we could achieve improvement in both travel time and emission at the same time? The experiment and simulation show that (1) a careful design of cost function with survey-supported cost coefficients would provide a solution to an optimized routing in multi-criteria traffic management (either at EU or SO), and (2) the network total travel time and emission could be effectively reduced when a routing strategy is well tuned and the emission is used as feedback in the system's control loop. © 2014 IEEE