碩士 / 淡江大學 / 土木工程學系 / 84 / Since the demand for travel are derived from the needs for
activities, some studies in 1970s had adopted activity-based
analysis approach to explore the course of travel-activity
behavior. Motivated by the need of having more understanding to
complex travel-activity behavior, this study incorporates with
activity characteristics to construct travel travel demand
models. Activity pattern groups can be defined as a set of
individuals with the similar travel-activity pattern
characterized by activity type, length of activity time, travel
time, modal choices and destination. The purpose of clustering
activity pattern groups, similar to market segmentation, is to
improve performance of trip prediction. In this study, the
activity patterns is classified into distinct groups by the
application of clustering techniques.According to the survey
from individuals in Taipei metropolitan area, the activity
pattern groups can be clustered into 3 groups. In addition,
discriminant analysis is applied to find discriminant functions
for the prediction of group membership for new samples. In
most of the previous studies, predictions of trip generation and
the modal choice were separately formulated in urban
transportation modeling system. However, the choice of
transportaton modes is also influenced by the choices of
activity types and activity location, this study seeks to
explore the interaction among the choices of activities,
destinations and transportation modes by conducting a
comprehensive survey on alternative activity, destination, and
mode is included in the formulation of utility functions to
capture the joint nature of choices. Finally, the demand models
are estimated by the method of nested Logit. Three different
multidimensional choices nested Logit models are established in
this study. The result of paramemter estimation shows that
nested Logit model with 3-dimensional choices represents a
better performance for individual''s travel-activity behavior,
that is, individual''s activity, destination and modal choice
behavior can be formulated as a sequential decision process.
Identifer | oai:union.ndltd.org:TW/084TKU00015019 |
Date | January 1996 |
Creators | Chen, Pey-Fen, 陳佩棻 |
Contributors | Shyr Feng-Yen, 石豐宇 |
Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
Language | zh-TW |
Detected Language | English |
Type | 學位論文 ; thesis |
Format | 104 |
Page generated in 0.0076 seconds