1991 Volume 9 Pages 101-108
This paper presents an estimation method of the number of visitors based on the destination-based trio surveys. The estimation method consists of the interrelated two steps: 1) estimating the most probable states of the number of visitors by routes based on traffic count data; 2) estimating the route choice models with choice-based surveys on visitors' behavior. We propose an analytical framework which can integrate the sub-models into a Stackelberg-type of programming model. The proposed methods can be considered as a logical extension of Cosslett's method for estimating logit models with general choice-based samples. The applicability of the methodology is verified by a case study conducted for the San'in Region. Simulation experiments are made to assess the impacts of survey errors on the reproductivity of the models.