Travel demand modeling for developing countries

Satyakala Jarugumilli, University of Nevada, Las Vegas

Abstract

The transportation planning process or travel demand forecasting process which is most commonly used at present, called the Urban Transportation Planning Process (UTPP), has its origins in studies performed about three decades ago in urban areas in the United States. This conventional planning process was designed primarily for the western/developed world where affordability and accessibility to a personal automobile is a common feature; The UTPP based models used in forecasting future travel demand rely heavily on socio-economic variables that are associated with an individual's propensity to travel. The socio-economic environments in the developing and the under-developed world are different from that prevailing in the developed world. Many urban areas in the third-world contain a mix of ethnic groups or households with structural differences in income and social characteristics. Thus, a direct application of the methodology of the UTPP may not be relevant to developing countries in third world settings; The main objectives of this research are to study the travel characteristics and sensitivity of key factors affecting travel demand of the people in the third world and to develop a framework that allows for the treatment of different groups based on one or more socio-economic features, such as income, household size etc; Trip generation models are developed to study the travel behaviour in a third world city, Dar-es-Salaam, capital of Tanzania, located in Africa. The study area is divided into 33 traffic analysis zones, and data for 22 zones of the 33 were available for this study. Typical Downtown, Residential, and Industrial areas are chosen for study. The influence of key trip generation socio-economic variables such as income, household size, age etc., are evaluated, and models developed for different groups under consideration. A marked variation is observed in the behavior of people depending upon where they reside. A single model for the entire study area proved to be unsatisfactory, thus highlighting some of the short-comings of the conventional process. Therefore, several models were calibrated. The travel behavior of individuals in the medium income group is less predictable. High income groups show a definite behavior. Also, recommendations are made to develop strategies to improve data collection procedures, in order to obtain accurate household information. Walk is a major mode of transport and thus, requires serious consideration in terms of mode choice calibration. Improvements in the operations of transit facilities and services, frequency, travel fare are likely to affect the travel behavior of the poor to a large extent. Other important variables not included in the study, such as distance to the downtown area, labor force in the household, ethnic origin etc., may prove to be significant in better explaining and forecasting travel demand. The terms "third-world" and "developing countries" are used interchangeably throughout the work.