Estimation of freight O-D matrix using Waybill data and traffic counts in Iran roads

Abstract

In this research, a realistic methodology, in order to estimate a true O-D matrix for freight movement in a designated region, is presented. This means, having had an initial O-D matrix (exp. Waybill data) and traffic count in many intercity roads, one can estimate an O-D matrix, having minimum distance from the initial matrix, and reproducing the observed traffic by assigning to the network. In the first step, after conducting studies and field work, some special route choice patterns in an intercity network were recognized by heavy vehicle drivers. Using the Logit Formulation, some probable (pure stochastic) assignment models related to the condition of the case were developed. Also, due to existing errors in both information sources, waybill and traffic count information, after recognition and determination of errors amount and their precision, based on Analytical Hierarchy Process Method, reliability of the mentioned data sources was estimated. Hence, to approach the best model of O-D matrix estimation, an Entropy Maximization model with a composed objective function was calibrated. Thus, having a calibrated matrix estimation model and a developed a traffic assignment model as a sub-model, true freight O-D matrix could be estimate.