Decomposition algorithm for statistical estimation of OD matrix with random link choice proportions from traffic counts

Statistical models for the estimation of Origin-Destination (OD) matrix fro m traffic counts that consider explicitly the presence of randomness in the link choice proportions have been developed recently. These models are mor e receptive to the fluctuations in the observations due to measurement erro rs and temporal variations and they can make better use of traffic informat ion. However, the estimation based on the new models involves the optimizat ion of functions that may not be convex and for large networks in real situ ations, conventional numerical algorithms such as Newton types may have dif ficulty in attaining the global optimum. In this paper, a decomposition alg orithm that makes combined use of the Coordinate Descent method and the Par tial Linearization Algorithm is proposed and its convergence proved. The pr oposed algorithm is shown to perform better with regard to finding the glob al optimum than the conventional quasi-Newton algorithm. Its implementation is demonstrated by a numerical example and a Hong Kong case study. (C) 199 9 Published by Elsevier Science Ltd. All rights reserved.