ABSTRACT
Parking policy can have very a strong and definitive effect on an individuals’ trip making characteristics and is thus considered as a major constituent of urban transport planning and policy. Control of the availability and use of parking spaces is an effective way of restricting the amount of traffic in central areas thus achieving wider social, economic and environmental goals. The design and implementation of such management plans requires the development of a predictor of car park occupancy. This chapter presents a method that is able to arrive at such predictions based on easily available data. Artificial neural networks appear to be efficient techniques for supporting complex parking management decisions. Finally, innovative solutions were given for the creation of both sets of models; A new clustering algorithm was constructed comparing Self-Organizing Feature Maps (SOFMs) instead of mean points and using an entropy related distance for carrying out the partitioning.
