ALGORITHMS FOR THE SOLUTION OF THE CONGESTED TRIP MATRIX ESTIMATION PROBLEM

Abstract

The problem of estimating trip matrices from traffic counts on a set of road links on congested networks is examined in this paper. The purpose is to develop efficient algorithms for the solution of the combined trip matrix estimation and User Equilibrium (UE) assignment problem. Two types of solutions are considered: the mutually consistent solution at which the two sub-problems are solved simultaneously; and the solution to the bi-level programming problem in which matrix estimation is the upper-level problem and UE assignment the lower-level problem. The algorithms are tested on two artificial networks and the Sioux Falls network. The algorithms developed are also applied to the combined signal optimisation and UE assignment problem.