Some remarks on macroscopic traffic flow modelling

This note contributes to the discussion on the current state and the prospectives of macroscopic traffic flow modelling. It is argued that macroscopic traffic flow models may never reach the descriptive accuracy level of other domains in physics and engineering. On the other hand, there is an eminent need for computer codes emulating dynamic traffic phenomena for various traffic engineering tasks (simulation, planning, control strategy design and testing etc.) hence the need to emphasize practical model validation work at the current stage of developments. Some properties of macroscopic traffic flow models are outlined and the usefulness of second-order approaches, as evidenced in empirical validation results, is underlined. (C) 1998 Elsevier Science Ltd. All rights reserved.