Iterative update of route choice proportions in OD estimation Available to Purchase

The origin–destination (OD) matrix, as a fundamental element of transportation planning, can be estimated in various ways. Owing to lower costs and the ability to keep matrices up to date, estimating OD matrices from observed traffic counts is gaining interest. These methods find an OD matrix that, when assigned to the network, reproduces the traffic counts. The main weakness of these methods, assuming traffic counts as deterministic values, has been overcome by the TFlowFuzzy (TFF) algorithm, which considers the counts as imprecise values. In this paper, a limitation of TFF is identified and accounted for with a methodological enhancement for modification. A key element in the estimation of trip matrices from traffic counts is the route choice proportions (assignment matrix). In the conventional TFF, the assignment matrix is computed only once and held constant during the whole estimation process. Hence, a modified algorithm is proposed in which the assignment matrix is updated successively. This modification is implemented on the real-size network of Mashad city. Results indicate a considerable improvement in the modified TFF goodness of fit compared to the traditional algorithm, as it is increased by 55%. This shows the effectiveness of the proposed algorithm, considering the assignment matrix endogenously.