| Summary: | 碩士 === 淡江大學 === 交通管理學系 === 86 === Probability or stochastic models have been widely applied to
address theuncertainty of traveler''s decision behavior. These
models were developed eitherin the field of mathematical
psychology or econometrics. In particular, therandom utility
theory has been widely implemented to study this sort ofbehavior
since 1970''s. However, after introduced by Zadeh in 1965, fuzzy
settheory has begun to gain attention to study the phenomenon of
ambiguous eventsrather than random nature, which may well be
suitable for addressing humandecision behavior such as traveling
decisions.This study is an attempt to apply fuzzy (set) theory
to study the auto drivingcommuters'' departure time and route
choice behavior. Under such framework,driver perceptions of
uncertain outcome of attributes affecting their routechoice is
due to the vagueness rather than randomness. A rule-based
reasoningprocess is therefore applied to model the observed
behavior rather thanthe commonly used utility maximization. Many
a recent studies have generallydemonstrate the possibility and
usefulness of similar applications but few hasshown physical
evidence.The drivers'' travel choice behavior is by its own
nature a complex issue, acontrolled experiment approach is
implemented in this study where a group ofselected auto-driving
commuters interact with a simulated commuting contextwith
repeated observations. The participants'' departure time and
route choicesare then recorded along with their individual
statement regarding their ownperception to the decision related
attributes.This study comprises of two stages. In the first
stage, each participant ispresented a preliminary survey
regarding his/her perception of choices androute (link)
attributes such as travel time,......,etc. In order to
investigatethe evidence that the "fuzziness" is indeed likely.
This include the bestdescription of how the decision on each day
was made routinely (i.e., thereasoning process and rule
aggregation ) and the perception of attributevariables ( the
membership function ). Further study in the second stage
isrepeated simulated commuting outcome with regard to
participants'' individualdeparture time and route choice
decisions. Repeated observations are presentedas consecutive
day-to-day manner to verify and calibrate the fuzzy
reasoningrule model while adopting membership functions and
decision rules from thefirst stage.
|
|---|
