Travel Demand with Forecasting Microsimulationby Konstadinos G. Goulias, Pennsylvania State Univ, University Park, United States,
Abstract: Transportation planners forecast travel demand using techniques that are based on dissagregate models estimated from cross-sectional data. Their forecasts are usually single future time point estimates. The use of cross-sectional models is based on the presumption that cross sectional variability in the sample is a valid indicator of changes over time. The input variables used by the usual travel demand models are obtained using approximate techniques. These procedures fail to capture effectively the relationships among socioeconomic, demographic, and travel behavior variables. The use of single time point forecasts is also an inefficient method of using the wealth and information available. The result is a questionable basis on which travel demand inferences are drawn. In this paper an alternative method designed for regional travel demand forecasting based on microsimulation and dynamic analysis and using data from the Netherlands is presented. This method combines the flexibility and richness of information of microsimulation with the realism and accuracy of the dynamic models of travel behavior. The forecasts are produced using a FORTRAN program (for the mainframe and personal computers). The program requires at least 15 MB of storage and in execution takes less than 30 minutes. The new system is composed by two components: (1) a socioeconomic and demographic component, which is providing forecasts for the numerous variables describing a 24 year life span of individuals and households; and (2) a mobility component, which is using the input from the first components to forecast household car ownership, trip generation, and modal split of households.
Subject Headings: Forecasting | Travel demand | Data processing | Computer models | Travel modes | Travel patterns | Cross sections | Computer software | Netherlands
Services: Buy this book/Buy this article
Return to search