Dear Sir/Madam



I am carrying out studies on traffic demand forecasting in the international scopes rather than domestic or urban ones. More precisely, I am looking for a model that is suitable to forecast both freight and people’s movement in the Mekong basin countries: i.e Cambodia, Lao PDR, Myanmar, and Thailand. Currently, they would like to develop highway networks that connect themselves together, and naturally, they would like to know how much traffic would use a certain road section in the network in the future.



As a researcher charging in this forecasting work, I reviewed various forecasting models and I realized that the conventional four- stage traffic forecasting model is not suitable to use for international transport at least for the present study areas. 



The four-stage forecasting model was originally devised to forecast traffic demand in the urban areas, and it requires O/D tables which are difficult to obtain or set up in the international scale. It is not that easy to get such specific statistics that can be utilized for developing O/D tables in the international scale.



However, I came across a study that was carried out in the similar fashion in the Central Asia region including Kazakhstan, Uzbekistan, Turkmenistan, Georgia, Armenia, Azerbaijan. In this study, typical four stage forecasting model (SATURN) was adopted. In fact, they developed a freight demand matrix after zoning the area in a very rough scale  (21 internal zones for the study area). They applied modal split rate to the obtained matrix to extract truck demand matrix. Then they assigned the demand onto the highway network that consists of highway routes of international importance. Finally, they validated their model by comparing estimated freight quantity with observed quantity in several selected links. (Actually, they converted number of trucks into the freight quantity in somehow)



I think I can apply this process to my study areas as well. But, many questions occurs in my mind such as what kind of link cost-flow function would be applied, what kind of time unit (a day, a month, or a year) will be employed for it, and how can we estimate traffic movement of passenger cars and buses?



I think I can obviously obtain macro socio-economic indicators like population, employment rate, imports/exports, number of registered vehicles etc. But I am not sure I can obtain traffic volumes for all road sections, although it is possible to obtain them partially. What shall I do with this information? Which model shall I adopt? What information I need to collect more for using this model?



If you have any idea or experience on developing an international traffic demand forecasting model, please give me an advice. Whatever it is, that will be very invaluable for me.





SANGJIN HAN (PhD in Transport Studies)



Department of Highway Research

Korea Transport Institute



2311, DAEHWA-DONG, ILSAN-GU, KOYANGSI

KYONGGIDO, KOREA, 411-701



Tel: +82-31-910-3112

Fax: +82-31-910-3235

E-mail:[log in to unmask]