A Study on Theory and Expansion of Present Pattern Method for OD Matrix Estimation

Abstract

In the conventional four-step transportation forecasting process is often used for modelling traffic demand in cities across Japan. In the trip distribution step, present pattern method is often used as it is served as an effective origin-destination (OD) estimation methodology when it is assumed that there are no significant changes on the connections between regions in the future and matches the marginal distribution of the traffic flow when the future trip production and trip attraction amount are known, while maintaining the distribution pattern of the current OD matrix. While empirical methods such as average factor method, Detroit method and Fratar method are well known, this article summarize the observation of present pattern methods from the point of view mathematical optimization and organize the theoretical foundations of each method. In addition, we discuss the possibility the possibility of expanding the present pattern method algorithm to estimate OD matrices that match aggregate traffic volumes other than trip production and trip attraction amount.