Data-driven design of on-demand transit service데이터 기반의 수요 대응형 교통 서비스 디자인

Cited 0 time in webofscience Cited 0 time in scopus
  • Hit : 195
  • Download : 0
On-demand transit service is one of the flexible transit services that are designed to update service routes in real-time according to passengers' requests. With the recent rapid development of autonomous driving, autonomous on-demand transit services are receiving lots of attentions. In recent years, unlike earlier studies that had focused on operational efficiency, many studies have dealt with the issue of service unreliability due to real-time route variability. To solve the service reliability issue, this dissertation focuses on two challenges: characterization of urban mobility patterns and on-demand transit service design. For the first challenge, this dissertation proposes a major mobility pattern extraction method using principal component analysis and a demand classification method based on spatiotemporal demand distributions. The proposed mobility pattern extraction method uses the daily consistency as a criterion for extracting major mobility patterns to be implemented into transportation system design. The major mobility patterns extracted based on the daily consistency can prevent the on-demand transit service from deteriorating service reliability due to random demands. In addition, a demand classification method for designing on-demand transit service is proposed. Since the demand data are classified based on the origin, destination, and departure time, it is possible to distinguish mobility data that are spatiotemporally clustered. Using the proposed classification method, the differences in the mobility patterns between the two regions can be easily identified. In addition, the distribution of the spatiotemporally grouped data can be applied to designing on-demand transit services. For the second challenge, this dissertation proposes a data-driven methodology for designing on-demand transit services and simulation studies to derive an optimal strategy for on-demand transit service. The proposed on-demand transit aims to improve service reliability with high operational efficiency. The on-demand transit system matches different types of on-demand transits services for different mobility patterns. For requests belonging to mobility patterns with high spatiotemporal density and high daily consistency, the flexibility of route operation is reduced and service reliability is increased. Other types of requests are matched to fully flexible route services that are designed to serve passengers whose origins and destinations are located in a wider space and time. The proposed on-demand transit service is validated through simulation studies. From the simulation results, the proposed service outperforms the existing on-demand transit system. In addition, through simulation studies based on various operation scenarios, optimal parameters for vehicle capacity and fleet size are derived, thereby improving operational efficiency and service quality at the same time.
Advisors
Yeo, Hwasooresearcher여화수researcher
Description
한국과학기술원 :건설및환경공학과,
Publisher
한국과학기술원
Issue Date
2021
Identifier
325007
Language
eng
Description

학위논문(박사) - 한국과학기술원 : 건설및환경공학과, 2021.8,[v, 123 p. :]

Keywords

on-demand transit service▼amajor mobility pattern▼aspatiotemporal demand classification▼aroute planning model▼areal-time route update; 수요 대응형 교통 서비스▼a주요 모빌리티 패턴▼a시공간적 수요 분류▼a경로 계획 모형▼a실시간 경로 업데이트

URI
http://hdl.handle.net/10203/294840
Link
http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=962575&flag=dissertation
Appears in Collection
CE-Theses_Ph.D.(박사논문)
Files in This Item
There are no files associated with this item.

qr_code

  • mendeley

    citeulike


rss_1.0 rss_2.0 atom_1.0