Development of transport models to predict flow-patiern and volume of traffic on the national highway network

Abstract

A transport modelling system has heen developed in this study to forecast traffic flow-pattern and traffic volume on the national highway network of Bangladesh. The model and the descrihed procedure are intended to provide highway planners with a tool for simple, fast and inexpensive estimation of traffic projections. This study attempts to comhine the transport models of trip distrihution and traffic assignment for forecasting traffic flow-pattern between old 20 districts and volume on the national highway network of Bangladesh. A douhly constrained gravity model of trip distribution and a capacity restraint traffic assignment model have heen computerized for this purpose. It is assumed that, the shortest path will be selected on the basis of minimum travel time and minimum travel cost for passenger and freight movements respectively. The gravity model is calibrated with the hase year (1990) data and the results of calibration are used to formulate regression equation of power form between travel resistance and the resistance variables lime and cost. The results of regression show that, the exponent for time is 2.05 with a correlation coefficient (r) of 0;98 and the exponent of cost is 1.76 with a correlation coefficient of 0.83. These higb values of correlation coefficient indicate that, the gravity model developed in this study can he used for reliable forecasting of traffic flow-pattern between old 20 districts of Bangladesh. The traffic assignment model has heen applied for forecasting traffic volume on the national highway network of Bangladesh. Growth rate of traffic for passenger and freight movements have heen calculated for all the 20 districts of Bangladesh assuming that, traffic growth rate is a function of some economic and demographic factors. Traffic volume on the road sections of the national highway network of Bangladesh has been projected for 20 years (1990-2010) at an interval of 5 years under existing condition of the network. The model has ,also heen applied to forecast traffic volume over Jamuna hridge in 1997 which is assumed to he'the opening year of the Jamuna hridge. To assess the validity of Forecasting, traffic volume over the Jamuna hridge in 1994 from this study and in the Jamuna hridge study have he en compared. This comparison shows a variation of only 5.5% for total' traffic (passenger and freight) between the predictions made in these two studies. The results also show that, due to the construction of the Jamuna hridge one of the heavily loaded Dhaka- Aricha highway will lose more than half of its traffic in 1997.