Stefano Manzo

Effects of Uncertainty in Speed–Flow Curve Parameters on a Large-Scale Model: Case Study of the Danish National Model

Uncertainty is inherent in transport models and prevents the use of a deterministic approach when traffic is modeled. Quantifying uncertainty thus becomes an indispensable step to produce a more informative and reliable output of transport models. In traffic assignment models, volume-delay functions express travel time as a function of traffic flows and the theoretical capacity of the modeled facility. The U.S. Bureau of Public Roads (BPR) formula is one of the most extensively applied volume delay functions in practice. This study investigated uncertainty in the BPR parameters. Initially, BPR parameters were estimated by analyzing observed traffic data related to the Danish highway network. Then, BPR parameter distributions were generated by using the resampling bootstrap technique. Finally, the generated parameter vectors were used to implement sensitivity tests on the four-stage Danish national transport model. The results clearly highlight the importance to modeling purposes of taking into account BPR formula parameter uncertainty, expressed as a distribution of values rather than assumed point values. Indeed, the model output demonstrates a noticeable sensitivity to parameter uncertainty. This aspect is evident particularly for stretches of the network with a high number of competing routes. Model sensitivity was also tested for BPR parameter uncertainty combined with link capacity uncertainty. The resultant increase in model sensitivity demonstrates even further the importance of implementing uncertainty analysis as part of a robust transport modeling process.

Effects of Uncertainty in Speed-Flow Curve Parameters on a Large-Scale Model

Uncertainty is inherent in transport models and prevents the use of a deterministic approach when traffic is modeled. Quantifying uncertainty thus becomes an indispensable step to produce a more informative and reliable output of transport models. In traffic assignment models, volume-delay functions express travel time as a function of traffic flows and the theoretical capacity of the modeled facility. The U.S. Bureau of Public Roads (BPR) formula is one of the most extensively applied volume delay functions in practice. This study investigated uncertainty in the BPR parameters. Initially, BPR parameters were estimated by analyzing observed traffic data related to the Danish highway network. Then, BPR parameter distributions were generated by using the resampling bootstrap technique. Finally, the generated parameter vectors were used to implement sensitivity tests on the four-stage Danish national transport model. The results clearly highlight the importance to modeling purposes of taking into account BPR formula parameter uncertainty, expressed as a distribution of values rather than assumed point values. Indeed, the model output demonstrates a noticeable sensitivity to parameter uncertainty. This aspect is evident particularly for stretches of the network with a high number of competing routes. Model sensitivity was also tested for BPR parameter uncertainty combined with link capacity uncertainty. The resultant increase in model sensitivity demonstrates even further the importance of implementing uncertainty analysis as part of a robust transport modeling process.

The effects of uncertainty in speed-flow curve parameters on a large-scale model: The Danish National Model case study

Uncertainty is inherent in transport models and prevents the use of a deterministic approach when traffic is modeled. Quantifying uncertainty thus becomes an indispensable step to produce a more informative and reliable output of transport models. In traffic assignment models, volume-delay functions express travel time as a function of traffic flows and the theoretical capacity of the modeled facility. The U.S. Bureau of Public Roads (BPR) formula is one of the most extensively applied volume delay functions in practice. This study investigated uncertainty in the BPR parameters. Initially, BPR parameters were estimated by analyzing observed traffic data related to the Danish highway network. Then, BPR parameter distributions were generated by using the resampling bootstrap technique. Finally, the generated parameter vectors were used to implement sensitivity tests on the four-stage Danish national transport model. The results clearly highlight the importance to modeling purposes of taking into account BPR formula parameter uncertainty, expressed as a distribution of values rather than assumed point values. Indeed, the model output demonstrates a noticeable sensitivity to parameter uncertainty. This aspect is evident particularly for stretches of the network with a high number of competing routes. Model sensitivity was also tested for BPR parameter uncertainty combined with link capacity uncertainty. The resultant increase in model sensitivity demonstrates even further the importance of implementing uncertainty analysis as part of a robust transport modeling process.