서브메뉴
검색
Data-Driven Population Tracking and Arrival Forecasting in Service Systems
Data-Driven Population Tracking and Arrival Forecasting in Service Systems
상세정보
- 자료유형
- 학위논문 서양
- 최종처리일시
- 20260202103120
- ISBN
- 9798315714682
- DDC
- 310
- 저자명
- Wood, Morgan.
- 서명/저자
- Data-Driven Population Tracking and Arrival Forecasting in Service Systems
- 발행사항
- [Sl] : The University of North Carolina at Chapel Hill, 2025
- 발행사항
- Ann Arbor : ProQuest Dissertations & Theses, 2025
- 형태사항
- 152 p
- 주기사항
- Source: Dissertations Abstracts International, Volume: 86-11, Section: B.
- 주기사항
- Advisor: Mersereau, Adam;Ziya, Serhan.
- 학위논문주기
- Thesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2025.
- 초록/해제
- 요약Motivated by applications at the Raleigh-Durham International Airport (RDU), this dissertation focuses on two objectives in service systems: real-time tracking of system population and forecasting future arrivals.The first objective involves estimating the number of passengers in the system based on noisy counts of arrivals and departures gathered from infrared-beam people counters. The noisiness of the data introduces challenges including the accumulation of errors over time. We develop policies that address these issues in population tracking, offering asymptotically optimal performance by periodically resetting population estimates based on a moving average of recent system departures. Additionally, we explore the impact of periodic inspections at a cost, showing that they offer greater flexibility in correcting errors and improved performance.The second objective of this dissertation involves forecasting future passenger arrivals, particularly in systems where arrivals are tied to scheduled events, such as flights. We introduce the concept of the "show-up profile," a parametric distribution modeling how early passengers arrive for their flights. These profiles are combined with a known flight schedule to generate aggregate passenger arrival forecasts. We develop a method for estimating show-up profiles using sensor data without the ability to map passengers to specific flights by considering a tractable approximation to the maximum likelihood that yields consistent estimates. Our forecasting approach, when compared to machine learning techniques, has comparable performance while offering greater interpretability and flexibility.This dissertation contributes to the application of Internet of Things (IoT) technology in service operations, particularly in environments with inaccurate or incomplete data. By integrating sensor-based data within policies and forecasting models, we demonstrate the value of IoT in improving operational decision-making in airports and other service systems.
- 일반주제명
- Statistics
- 일반주제명
- Information technology
- 키워드
- Queues
- 키워드
- Machine learning
- 기타저자
- The University of North Carolina at Chapel Hill Statistics and Operations Research
- 기본자료저록
- Dissertations Abstracts International. 86-11B.
- 전자적 위치 및 접속
- 로그인 후 원문을 볼 수 있습니다.
MARC
008260126s2025 us c eng d■001000017357033
■00520260202103120
■006m o d
■007cr#unu||||||||
■020 ▼a9798315714682
■035 ▼a(MiAaPQ)AAI31937857
■040 ▼aMiAaPQ▼cMiAaPQ
■0820 ▼a310
■1001 ▼aWood, Morgan.
■24510▼aData-Driven Population Tracking and Arrival Forecasting in Service Systems
■260 ▼a[Sl]▼bThe University of North Carolina at Chapel Hill▼c2025
■260 1▼aAnn Arbor▼bProQuest Dissertations & Theses▼c2025
■300 ▼a152 p
■500 ▼aSource: Dissertations Abstracts International, Volume: 86-11, Section: B.
■500 ▼aAdvisor: Mersereau, Adam;Ziya, Serhan.
■5021 ▼aThesis (Ph.D.)--The University of North Carolina at Chapel Hill, 2025.
■520 ▼aMotivated by applications at the Raleigh-Durham International Airport (RDU), this dissertation focuses on two objectives in service systems: real-time tracking of system population and forecasting future arrivals.The first objective involves estimating the number of passengers in the system based on noisy counts of arrivals and departures gathered from infrared-beam people counters. The noisiness of the data introduces challenges including the accumulation of errors over time. We develop policies that address these issues in population tracking, offering asymptotically optimal performance by periodically resetting population estimates based on a moving average of recent system departures. Additionally, we explore the impact of periodic inspections at a cost, showing that they offer greater flexibility in correcting errors and improved performance.The second objective of this dissertation involves forecasting future passenger arrivals, particularly in systems where arrivals are tied to scheduled events, such as flights. We introduce the concept of the "show-up profile," a parametric distribution modeling how early passengers arrive for their flights. These profiles are combined with a known flight schedule to generate aggregate passenger arrival forecasts. We develop a method for estimating show-up profiles using sensor data without the ability to map passengers to specific flights by considering a tractable approximation to the maximum likelihood that yields consistent estimates. Our forecasting approach, when compared to machine learning techniques, has comparable performance while offering greater interpretability and flexibility.This dissertation contributes to the application of Internet of Things (IoT) technology in service operations, particularly in environments with inaccurate or incomplete data. By integrating sensor-based data within policies and forecasting models, we demonstrate the value of IoT in improving operational decision-making in airports and other service systems.
■590 ▼aSchool code: 0153.
■650 4▼aStatistics
■650 4▼aInformation technology
■653 ▼aInternet of Things
■653 ▼aOperations management
■653 ▼aParametric learning
■653 ▼aQueues
■653 ▼aMachine learning
■690 ▼a0454
■690 ▼a0796
■690 ▼a0463
■690 ▼a0489
■71020▼aThe University of North Carolina at Chapel Hill▼bStatistics and Operations Research.
■7730 ▼tDissertations Abstracts International▼g86-11B.
■790 ▼a0153
■791 ▼aPh.D.
■792 ▼a2025
■793 ▼aEnglish
■85640▼uhttp://www.riss.kr/pdu/ddodLink.do?id=T17357033▼nKERIS▼z이 자료의 원문은 한국교육학술정보원에서 제공합니다.
