Urban road network modeling and real-time prediction based on householder transformation and adjacent vector

Abstract

This paper put forward a multivariate one-order-regression single road link model based on the algorithm of Householder Transformation to reduce the computation complexity in real-time prediction and to facilitate the study on network turn-ratio pattern evolution. Then the paper analyses the limitation of current urban road network model based on adjacent matrix and contributed a novel model based on new memory strategy aiming at reduce the memory space occupied by adjacent matrix, carrying turn movement information in the storage and avoiding redundant calculation. To verify the new modeling method, the study involved in a field work on part of urban network in Beijing, China. In conclusion, the new modeling methods in this paper enhanced the performance of urban road modeling. © 2009 Springer Berlin Heidelberg.