서브메뉴
검색
Shared-Use Autonomous Mobility Systems: Planning, Pricing and Mixed-Service Operation
Shared-Use Autonomous Mobility Systems: Planning, Pricing and Mixed-Service Operation
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20250211151254
- ISBN
- 9798381977080
- DDC
- 385
- 저자명
- Abkarian, Hoseb.
- 서명/저자
- Shared-Use Autonomous Mobility Systems: Planning, Pricing and Mixed-Service Operation
- 발행사항
- [Sl] : Northwestern University, 2024
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2024
- 형태사항
- 182 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 85-10, Section: B.
- 주기사항
- Advisor: Mahmassani, Hani.
- 학위논문주기
- Thesis (Ph.D.)--Northwestern University, 2024.
- 초록/해제
- 요약Autonomous vehicle (AV) technology will bring about significant changes to both transportation systems and human behavior. One transportation service that may significantly benefit from AVs is ridesharing, since relying on human drivers can be both expensive and difficult to manage. Hence, in the near future, it is likely that Transportation Network Companies (TNCs) will own and operate AVs in a shared-use autonomous mobility system (SAMS).As a result, several ridesharing business models may appear. For example, unlike current carsharing models like Zipcar, users will not need to walk to reach the autonomous vehicle nor worry about finding parking spots. Consequently, AVs have the potential to greatly enhance the viability of this business model. Additionally, as AV technology becomes prevalent, users' preferences might shift towards specific service attributes. For instance, they might increasingly opt for reserving hourly-based services in advance. Users could also become less concerned about deviations to pick up other riders or packages during the trip, as they may be more productive in AVs (working, reading, etc.).From an operational perspective, operators of AV fleets will have comprehensive information about and control over their vehicles. This paves the way for efficient fleet management strategies that need to be studied for new business models. Furthermore, AV fleet operators will likely maintain a constant fleet size, escaping the restriction of human drivers who are limited by their working hours, and hence, AVs could be utilized for a significant portion of the day.Considering the above, and given the diverse spatial and temporal demand patterns for various services, it will be beneficial for operators to make AVs versatile, allowing them to serve different purposes to maximize utilization, profitability and/or societal benefits.Therefore, this thesis tackles extending the body of knowledge on SAMS by studying new business models that could emerge in the near-to-medium future. Specifically, this thesis focuses on the concept of mixed-service operation, whereby vehicle fleets are pooled to conduct different purpose trips, rather than having separate fleets. This thesis also tackles another important aspect that is important for these TNCs, which is the ability to price rides in real time. Given that customer behavior can change either quickly (e.g. pandemic) or slowly (e.g. introduction of another service), pricing algorithms should be able to adapt their algorithms to quickly optimize the system goal. This will lead to TNCs requiring novel pricing techniques.Chapter 3 studies a mixed-service operation of SAMS where customers can request rides either immediately or through reservations and use the AV for a point-to-point service or a time-slot-based rental service, respectively. Three autonomous-vehicle-to-user assignment strategies are presented, a first-come-first-serve strategy and two optimization-based (bipartite matching) strategies. The mathematical formulations attempt to achieve a good tradeoff between the wait times of reservation-based users and on-demand users, while minimizing overall empty fleet miles. A case study in Chicago is presented using taxi data, and the combined mixed-service fleet operation is compared to a case with two separate operations: one for an on-demand point-to-point service and one for a reservation-based time-slot rental service. Results show that a combined mixed-service operation can provide a more balanced service than the case with two separate operations with respect to the key performance measures of wait time and empty fleet miles.Chapter 4 takes a holistic approach to modeling SAMS, while specifically focusing on novel pricing techniques using reinforcement learning (RL). Most of the literature lacks in building realistic simulators of SAMS, and so, this chapter focuses on including most important aspects of ridesharing to produce reliable results, including vehicle assignment, customer choice process, wait time modeling, and pricing, with pricing being the main focus of this chapter.
- 일반주제명
- Transportation
- 기타저자
- Northwestern University Civil and Environmental Engineering
- 기본자료저록
- Dissertations Abstracts International. 85-10B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008250123s2024 us c eng d■001000017161082
■00520250211151254
■006m o d
■007cr#unu||||||||
■020 ▼a9798381977080
■035 ▼a(MiAaPQ)AAI30991031
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a385
■1001 ▼aAbkarian, Hoseb.▼0(orcid)0000-0003-2681-3002
■24510▼aShared-Use Autonomous Mobility Systems: Planning, Pricing and Mixed-Service Operation
■260 ▼a[Sl]▼bNorthwestern University▼c2024
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2024
■300 ▼a182 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 85-10, Section: B.
■500 ▼aAdvisor: Mahmassani, Hani.
■5021 ▼aThesis (Ph.D.)--Northwestern University, 2024.
■520 ▼aAutonomous vehicle (AV) technology will bring about significant changes to both transportation systems and human behavior. One transportation service that may significantly benefit from AVs is ridesharing, since relying on human drivers can be both expensive and difficult to manage. Hence, in the near future, it is likely that Transportation Network Companies (TNCs) will own and operate AVs in a shared-use autonomous mobility system (SAMS).As a result, several ridesharing business models may appear. For example, unlike current carsharing models like Zipcar, users will not need to walk to reach the autonomous vehicle nor worry about finding parking spots. Consequently, AVs have the potential to greatly enhance the viability of this business model. Additionally, as AV technology becomes prevalent, users' preferences might shift towards specific service attributes. For instance, they might increasingly opt for reserving hourly-based services in advance. Users could also become less concerned about deviations to pick up other riders or packages during the trip, as they may be more productive in AVs (working, reading, etc.).From an operational perspective, operators of AV fleets will have comprehensive information about and control over their vehicles. This paves the way for efficient fleet management strategies that need to be studied for new business models. Furthermore, AV fleet operators will likely maintain a constant fleet size, escaping the restriction of human drivers who are limited by their working hours, and hence, AVs could be utilized for a significant portion of the day.Considering the above, and given the diverse spatial and temporal demand patterns for various services, it will be beneficial for operators to make AVs versatile, allowing them to serve different purposes to maximize utilization, profitability and/or societal benefits.Therefore, this thesis tackles extending the body of knowledge on SAMS by studying new business models that could emerge in the near-to-medium future. Specifically, this thesis focuses on the concept of mixed-service operation, whereby vehicle fleets are pooled to conduct different purpose trips, rather than having separate fleets. This thesis also tackles another important aspect that is important for these TNCs, which is the ability to price rides in real time. Given that customer behavior can change either quickly (e.g. pandemic) or slowly (e.g. introduction of another service), pricing algorithms should be able to adapt their algorithms to quickly optimize the system goal. This will lead to TNCs requiring novel pricing techniques.Chapter 3 studies a mixed-service operation of SAMS where customers can request rides either immediately or through reservations and use the AV for a point-to-point service or a time-slot-based rental service, respectively. Three autonomous-vehicle-to-user assignment strategies are presented, a first-come-first-serve strategy and two optimization-based (bipartite matching) strategies. The mathematical formulations attempt to achieve a good tradeoff between the wait times of reservation-based users and on-demand users, while minimizing overall empty fleet miles. A case study in Chicago is presented using taxi data, and the combined mixed-service fleet operation is compared to a case with two separate operations: one for an on-demand point-to-point service and one for a reservation-based time-slot rental service. Results show that a combined mixed-service operation can provide a more balanced service than the case with two separate operations with respect to the key performance measures of wait time and empty fleet miles.Chapter 4 takes a holistic approach to modeling SAMS, while specifically focusing on novel pricing techniques using reinforcement learning (RL). Most of the literature lacks in building realistic simulators of SAMS, and so, this chapter focuses on including most important aspects of ridesharing to produce reliable results, including vehicle assignment, customer choice process, wait time modeling, and pricing, with pricing being the main focus of this chapter.
■590 ▼aSchool code: 0163.
■650 4▼aTransportation
■650 4▼aEnvironmental engineering
■653 ▼aAutonomous vehicle
■653 ▼aTransportation systems
■653 ▼aTransportation Network Companies
■653 ▼aService operation
■690 ▼a0709
■690 ▼a0543
■690 ▼a0775
■71020▼aNorthwestern University▼bCivil and Environmental Engineering.
■7730 ▼tDissertations Abstracts International▼g85-10B.
■790 ▼a0163
■791 ▼aPh.D.
■792 ▼a2024
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17161082▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
