International Scientific Journals
of Scientific Technical Union of Mechanical Engineering "Industry 4.0"

In order to establish this model, several independent variables of aggregate level and dependent variables have been taken into account for traffic production of a region for a period of 24 hours. The objective of this paper is to identify and estimate the main variables which are significant for development of a suitable model for forecasting the traffic production for Anamorava region. Establishment and assessment of a model will be done using techniques such as Multiple Regression Analysis (MRA) and Radial Basis Function Neural Network (RBFNN). The results gained according to these two techniques will be compared to find out differences. In this regard performance analysis through indicators R², ME, RSME which are also used to show errors in forecasting. The results are in favour of RBFNN model because it performs better than MRA in forecasting traffic production.

The contemporary lifestyle leads to rapid increase of number of motorized traffic participants thus, there is a need of traffic planning. The task of any modern traffic planning is creating such road network that will be reviewed, safe, with appropriate capacity, economical in terms of exploitation costs and cost for building, also designed to be compatible with structure of the city. In the field of traffic engineering exist a wide range of software packages for different purposes, with different performance and quality. One of them is world leading software package for traffic analyzes and forecast of transport demand, PTV Vision VISUM.

In this paper will be represented the steps of the methodology for calculation of transport demand, starting from setting the background, drawing the traffic network until the forecast using an appropriate software package PTV Vision – VISUM in which the base is standard four step transport demand model.