Calibrating aggregate travel demand models with traffic counts: Estimators and statistical performance

摘要: Traffic counts on network links constitute an information source travel demand which is easy to collect, cheap and repeatable. Many models proposed in recent years deal with the use of traffic estimate Origin/Destination (O/D) trip matrices under different assumptions type "a-priori" available (surveys, outdated estimates, models, etc.) assignment mapping (see Cascetta & Nguyen 1988). Less attention has been paid possibility using parameters models. In this case most methods are relative particular model structures (e.g. gravity-type) statistical analysis estimator performance not thoroughly carried out. paper a general framework defining Maximum Likelihood, Non Linear Generalized Least Squares (NGLS) Bayes estimators aggregated combining counts-based other sources (sample or priori estimates) first, thus extending generalizing previous work by authors (Cascetta Russo 1992). Subsequently solution algorithm projected-gradient for NGLS given its convenient theoretical computational properties. The based combination analytical/numerical derivates order make applicable Statistical performances evaluated small test through Monte Carlo method repeatedly sampling "starting estimates" (known) generation/distribution/modal split/assignment system Tests were out assuming levels "quality" starting estimates numbers counts. Finally was applied calibration described real medium-size Italian town very satisfactory results terms both parameter values counted flows reproduction.