Updated Bureau of Public Roads (BPR) curves for freeways and signalised arterials are needed to improve the accuracy of speed estimates used in transportation demand models. These updated curves generally involved the use of higher power functions which show relatively little sensitivity to traffic volume changes until the demand exceeds capacity. The predicted speed drops abruptly to a very low value when the demand exceeds capacity. It has been demonstrated that the curves provide improved estimates of vehicle speeds under both uncongested and queuing conditions; however, the impact of these curves on the performance of travel demand models has not been investigated in the past. Practitioners have been concerned about the impacts of such abrupt speed-flow curves on the performance of their travel demand models. It has been stated that higher power functions are more difficult computationally for computers to evaluate and that more abrupt speed-flow curves adversely affect the rate of convergence to equilibrium solutions in the traffic assignment process. This paper investigates the impact of the updated speed-flow curves on the performance of selected travel demand models and demonstrates that an alternative speed-flow equation has similar or better accuracy and significantly superior convergence properties during the traffic assignment process. (a).
Abstract