In this paper, the authors present a Neural-Geo-Temporal Model (NGTM) that incorporates temporal interactions between transportation systems and land use patterns as the fundamental elements to express urban dynamics and its effects on travel demand. NGTM intends to reach an efficient representation of geographic, evolutionary and non-linear characteristics of urban interactions, without incurring on additional costs (data collection, processing time, personnel). NGTM combines Neural Networks (NN), which allows the creation of a mathematical model without considering a priori relations between dependent and independent variables, and Geographical Information System (GIS), which contributes on conducting spatial analysis that generate data/information on the evolution and characteristics of the urban area. This paper is divided into five sections. After this brief introduction, the authors present a brief review on NN fundamentals, which is followed by the theoretical conception of the NGTM, the case study and the conclusions. (Author/publisher) For the covering entry of this conference, please see ITRD abstract no. E210413.
Samenvatting