A NEURAL NETWORK MODEL TO EFFICIENTLY SIMULATE THE TRANSPORTATION SYSTEM

Abstract In recent years the attention of the analysts of transportation systems has been addressed to the search of new tools able to quickly simulate the response of the transportation system to modifications in its configuration. Neural Networks represent a new kind of information processing, their main attractions being the capability to work in real time, to process also incomplete and wrong data and above all to found a relationship among data when the underlying mathematical function is not expressed in a closed form. In this paper, a new kind of Neural Network is described, that can be identified in an electrical circuit; the information is then identified in an electrical current moving along the circuit that can represent the topological structure, if one, of the examined problem. Fot this reason this kind of Neural Network can be usefully used in the transportation field, by identifying the transportation network with the topological structure of the Neural Network.