Choice Model Refinement from Network Data

Abstract

This article considers the problem of refining the parameter s of disaggregate traveldemand models using observations of time-dependent network flows. The mea-surement function that links the behavioral model parameters to the network flowsis given through an iterated DTA (dynamic traffic assignment ) microsimulation.The challenging aspect of this configuration is that the meas urement equation isnot given in closed form but is the result of a complex simulation process, in thecourse of which disaggregate models of travel behavior and of network flows arerepeatedly evaluated until a state of consistency between demand and supply isattained (Nagel and Flotterod, forthcoming; Cascetta, 1989).The increased availability of detailed network surveillance data triggered recentefforts to calibrate behavioral model parameters (and also network supply param-eters) jointly with the origin/destination flows represent ing travel demand levels(Antoniou et al., 2007; Balakrishna, 2006; Vaze et al., 2009). These approaches,however, resort to black box optimization techniques that, by design, exploit prob-lem structure at most in terms of a numerical linearization. The approach of Flot-terod et al. (2011) is an exception; here, a tractable analytical approximation of acomplete DTA microsimulation system is developed and exploited. That research

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 — 5 citations exist upstream.

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