Analysis of the Evolution of Travel Demand in Urban Areas: a Neural-geo-temporal Modelling Approach

Considerable amount of research efforts has been dedicated to conceive forecasting models for predicting urban development and travel demand. Mainly originated from the application of economic theory, urban development models were initially conceived to explain the configuration and evolution of urban structures (Chapin, 1957). According to Ross (1988), these models are used to predict how future activities will be allocated, assuming that the land uses are determined by the location, the availability of services and proximity to other types of land uses. Based on these pioneer initiatives, Wegener (2003) identified and analyzed 20 integrated land use-transport models, which incorporated the most important spatial processes of development in conjunction with travel demand forecasting. Most part of them predicts land use separately to travel demand forecasting, i.e. each element (land use or transport) is modeled independently using its outputs for subsequent steps of the modelling process. Traditionally, planners have used these models to estimate future land use patterns and from them travel demand is estimated (Miller and Demestky, 1999), which is subsequently used to estimate traffic flows, congestion and pollution.