Using 'SATURN' to determine the effect of data on model accuracy and operation efficiency on a network of a small CBD

Abstract

The classic approach to the construction of a mathematical model to simulate traffic movements in a dense road network, such as a Central Business District (CBD), includes the need to prepare a complete Origin - Destination (O-D) matrix of vehicle trips into and through the network. Subsequent steps in the model deal with the coding of the road network and the assignment of the 0-D matrix onto the network. The acquisition of reliable origin - destination data to create the 0-D matrix is very often a tedious and costly exercise The use of traffic count information as an alternative method of 0-D data creation needs consideration. Using the Simulation and Assignment of Traffic in Urban Road Network (SATURN) suite of computer programs, the task of simulating the prevailing traffic conditions may be considerably facilitated and hence reducing the cost of a traffic study in a CBD area. The present project report deals with that particular issue in the context of a small CBD in South Africa. The results show that saving can be achieved in the data collection.