In recent years, the popularity of geographic information systems (GIS) has grown enormously across many disciplines and in numerous sectors. With this increased reliance on GIS and vastly improved computer speed and storage capabilities, considerable effort has been expended in increasing the accuracy of the GIS road networks around the world. This process can, however, be taken to extremes, delivering undetectable improvements in geographic accuracy at the cost of much degraded application performance. This paper describes the use of a minimum path approach to optimally reduce the size of a GIS road network within user-defined tolerance limits. The technique is easily implemented and fast to effect. Further, the minimum path algorithm has been adapted to deal with road networks in cities of any size. The paper also discusses the application of the technique to Melbourne's extensive road network and outlines some of the benefits that have already been delivered in GIS-based modelling applications at the Transport Research Centre, Royal Melbourne Institute of Technology (Melbourne, Australia). (a).
Abstract