Showing papers by "Margaret Bell published in 1992"

Network programming to derive turning movements from link flows

Abstract: It is generally accepted that there is a need to develop traffic models to manage and control congestion in real time. This control needs to be integrated with route guidance systems in order to achieve rerouting. A basic requirement of such models is to automatically establish turning movements from link flows. Conventional models such as entropy maximizing and information minimizing have been developed for use off line for transportation planning and are not suitable for application on line. A novel approach is proposed that uses linear programming to forecast traffic congestion in an urban network and define junction turning flows. The algorithms, originally developed for optimizing flows of water and electricity, use detector flow measurements, weighted links, and constrained upper and lower flow bounds. The principles underlying this approach are explained. The development, calibration, validation, and implementation of the model in a real network in the city of Leicester, England, are described. The results show that turning movements can be predicted with sufficient accuracy to justify further work, in particular to carry out a demonstration of the application of the algorithm on street.