知識利用型の経路選択モデル化手法

抄録

There are advanced techniques using idea of Knowledge Base (KB) to describe human behaviour. Three methods such as production system (PS), fuzzy reasoning/logic (FL) and neural networks (NN) are applied to modelling of binary route choice. The basic concept of each method is mentioned. The advantages and disadvantages of them are discussed from the comparison with logit modelling as a standard method for solving the problem. The difference of KB approach from statistical one is considered. And also the relation among these techniques are summarized. Finally, a combined model between FL and NN can be proposed to develop these methods in terms of accurate estimation and precise description of human travel behaviour.