| Summary: | 碩士 === 淡江大學 === 交通管理學系 === 87 === The dynamic travel choice behavior is by its own nature a complex issue. Probabilistic models have been widely applied to address the uncertainty of travel''''''''s decision behavior. In particularly, random utility theory has been implemented to study this sort of behavior since 1970''''''''s. While fuzzy set theory has begun to gain attention to handle the phenomenon of ambiguous events rather than random nature, which may well be suitable for addressing human decision behavior such as route choices.
Neuro fuzzy models are proposed as the combinations of fuzzy reasoning and neural network. Basically combination between fuzzy reasoning and neural network was proposed as an idea of fuzzy neuron. Well-known description with fuzzy congnitive map was proposed by Kosoko''''''''s works (1986, 1992). Many practical neuro fuzzy models have been introduced by Japanese researchers since 1980''''''''s. However these models were generally in fuzzy control context rather than travel behavioral ones. Few noteworthy studies were those by Akiyama 1997, Matsuura, 1997 and Tsuboi 1997, however none of them considered dynamic aspect of travel behavior.
In this proposed study, fuzzy neuro models are to be investigated for dynamic route choice behavior. Emphasizing the commuters’ dynamic decision behavior model. It is therefore the main purpose of the author to establish an in-depth investigation of the dynamic aspect of neuro-fuzzy networks in dynamic route choices through integration of well-established dynamic travel behavior framework by the author. The observation is conducted through controlled experiment manner considering the cost effectiveness of this approach..
|
|---|
