| Summary: | 碩士 === 淡江大學 === 交通管理學系 === 86 === Title of Thesis: Total Page:
133A Study of Schastic Dynamic Assignment Model with Fuzzy
Measure Traffic InformationName of Institute: Graduate Institute
of Transportation Science, Tamkang UniversityGraduate Data: June
1998 Degree Conferred : MasterName of Student:
Pao-Huey Huang Advisor: Dr. Shihsien LiuAbstract: In
a real road network, drivers make their route choice decisions
based on predicted future travel time and the outcome of route
choice decisions displays a stochastic distribution. This
behavior condition can be perfectly described by fuzzy measure
of dynamic travel time from Dubois''s invertible mapping of one
dimension route choice probability distribution. Hence, this
study focuses stochastic dynamic assignment model with fuzzy
measure traffic information. Using the indexes of necessary
and possible, we can not only improve the quality of model
performance, but also provide a new approach to tackle the
dynamic traffic assignment problem. The study area locates at
the Lin-Kou area, Taipei County, because this network is
isolated and traffic zones are separable. The license plate
survey and outbound traffic zone questionnaire survey are
performed simultaneously to capture the drivers'' perceived and
real route choice outcome. According to the study, the
results lead to some significant findings as follows:1. The
deterministic model is a special case of the stochastic model.2.
When route travel time with non-symmetric fuzzy membership
function, the likely shortest route may change with different
alpha-cut value.3. In the Lin-Kou area, the dynamic route choice
set among a pair of origin and destination is stable. The
reliable degree of perceived fuzzy membership function of route
travel time,alpha-cut value, is 0.4.4. The parameter of a
stochastic assignment model,theta , may fluctuate based on
location and time of the day. This make the
implementation of logit or probit based stochastic is
impracticable.Key works: Fuzzy measure, Invertible mapping,
Travel time fuzzy number
|
|---|
