The effectiveness of macroscopic dynamic freeway flow models at both uninterrupted and interrupted flow conditions is tested. Model implementation is made by finite difference methods developed here for solving the system's governing equations. These schemes are more effective than existing numerical methods, particularly when generation terms are introduced. The modelling alternatives and numerical solution algorithms are compared by employing a data base generated through microscopic simulation. Despite the effectiveness of the proposed numerical treatments, substantial deviations from the data at interrupted flows are still noticeable. In order to improve performance when flow is interrupted, we develop a modelling methodology that takes into account the ramp-freeway interactions so that all freeway components are treated as a system. We show that the coupling effects of the merging traffic streams are significant. Finally, the incremental benefits of using the more sophisticated high-order continuum models are assessed.
Bibliographical noteFunding Information:
This research was supported by NSF Grant CEE 8210189P. reliminaryt estinga nd validationw as perfor-medb y Mr. Y. Yamauchi,w ho is currentlyw ith theJ apan Highway Corporation,T okyo, Japan.
- free way flow
- freeway modelling
- macroscopic traffic flow dynamics
- traffic simulation