A Fuzzy-Neural Network Model for Travel Demand on Evacuee-Trip Production at the beginning of the Great Hanshin-Awaji Earthquake

Abstract

This paper describes applying the FNN (fuzzy-neural network)model concerning travel demand on evacuee-trip production at the aggregate level at the beginning of the Great Hanshin-Awaji Earthquake and compares it with functional models. The FNN model has been evaluated with good fitness upon application to travel demand: The determination coefficient of the FNN model and the functional models are 0.75 and less than 0.40,respectively. Furthermore, the FNN model demonstrates that the average number of family members, the difficulties in acquiring information, and the rate of collapsed buildings and / or houses are the major factors influencing the generation of travel demand of evacuee-trip production at the aggregate level.