Evaluation with Traffic Assignment Under Demand Uncertainty

Abstract

The impact of demand uncertainty on the evaluation of network improvements when using traffic assignment is addressed. Theoretical results indicate that the expected value of the performance of the traffic system is not necessarily equivalent to the performance of the system at the expected value of demand, and therefore the expected demand cannot be used with disregard of the variance in demand forecasts. Using expected demand tends to overestimate performance of the network and could lead to erroneous choice of improvements. Alternative possibilities for dealing with this problem are explored, including an approach in which the demand is inflated. This yield benefits not only in terms of selecting improvements with lower expected total system travel time but also significant reductions in the variance associated with this measure. Demand inflation should take place after a budget has been set, so that decisions resulting from inflation do not dictate that more infrastructure be built but that different improvements be selected. Furthermore, numerical examples indicate that there is a significant probability that the deterministic traffic assignment procedure will incorrectly rank improvement policies. This poses a serious concern for decisions based on this procedure, so potential actions are discussed to address the problem.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 96 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.