Consequences of stochastic assignment matrix upon transportation demand’s estimate by mean of traffic counts

This paper deals with the problem of estimating parameters of a transportation demand linear model from traffic counts: the assignment matrix is assumed to be stochastic. The study of asymptotic properties shows that the least squares estimator is not consistent. A lot of simulations on a test problem have been carried out: the properties of the least squares estimator, in the case considered, are revealed to be good, and are also good in the presence of the biased estimate of some parameters upon which the assignment matrix depends.