O–D MATRIX ADJUSTMENT FOR TRANSIT NETWORKS BY CONJUGATE GRADIENT ITERATIONS

Abstract

The adjustment of an obsolete demand matrix, from some given known data, is an important issue for transport research. In this article we introduce a penalized model, based on volume counts on a given set of arcs or segments, to update the demand matrix. Also, we propose a multiplicative conjugate gradient algorithm to solve the resultant convex optimization problem. This algorithm has been programmed with the macro language of EMME and tested with a synthetic scenario from the Winnipeg network. The numerical results show that the proposed algorithm improves the performance of the traditional multiplicative steepest descent algorithm, introduced by Spiess