Abstract
This paper proposes a method of inverse-estimating the unknown parameters of the composition model which composed of integrated network equilibrium model using observed traffic. A bi-level optimization problem is built in order to estimate unknown parameters of the [MSUE/ATIS-Demand] model and link cost function. It is proved that a bi-level optimization problem can be efficiently calculated by nonlinear sensitivity analysis. The following results about the estimation of the unknown parameters are found: (1) there is an effective setting method for a set of observed links; (2) it is possible to improve precision of estimate by making the observation error small.
