{"title":"A Time Series Decomposition and Reinforcement Learning Ensemble Method for Short-Term Passenger Flow Prediction in Urban Rail Transit","authors":"Jinxin Wu, Deqiang He, Xianwang Li, Suiqiu He, Qin Li, Chonghui Ren","doi":"10.1007/s40864-023-00205-1","DOIUrl":null,"url":null,"abstract":"<p>Short-term passenger flow prediction (STPFP) helps ease traffic congestion and optimize the allocation of rail transit resources. However, the nonlinear and nonstationary nature of passenger flow time series challenges STPFP. To address this issue, a hybrid model based on time series decomposition and reinforcement learning ensemble strategies is proposed. Firstly, the improved arithmetic optimization algorithm is constructed by adding sine chaotic mapping, a new dynamic boundary strategy, and adaptive T distribution mutations for optimizing variational mode decomposition (VMD) parameters. Then, the original passenger flow data containing nonlinear and nonstationary irregular changes of noise is decomposed into several intrinsic mode functions (IMFs) by using the optimized VMD technology, which reduces the time-varying complexity of passenger flow time series and improves predictability. Meanwhile, the IMFs are divided into different frequency series by fluctuation-based dispersion entropy, and diverse models are utilized to predict different frequency series. Finally, to avoid the cumulative error caused by the direct superposition of each IMF’s prediction result, reinforcement learning is adopted to ensemble the multiple models to acquire the multistep passenger flow prediction result. Experiments on four subway station passenger flow datasets proved that the prediction performance of the proposed method was better than all benchmark models. The excellent prediction effect of the proposed model has important guiding significance for evaluating the operation status of urban rail transit systems and improving the level of passenger service.</p>","PeriodicalId":44861,"journal":{"name":"Urban Rail Transit","volume":null,"pages":null},"PeriodicalIF":1.7000,"publicationDate":"2023-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Urban Rail Transit","FirstCategoryId":"5","ListUrlMain":"https://doi.org/10.1007/s40864-023-00205-1","RegionNum":4,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"TRANSPORTATION","Score":null,"Total":0}
引用次数: 0
Abstract
Short-term passenger flow prediction (STPFP) helps ease traffic congestion and optimize the allocation of rail transit resources. However, the nonlinear and nonstationary nature of passenger flow time series challenges STPFP. To address this issue, a hybrid model based on time series decomposition and reinforcement learning ensemble strategies is proposed. Firstly, the improved arithmetic optimization algorithm is constructed by adding sine chaotic mapping, a new dynamic boundary strategy, and adaptive T distribution mutations for optimizing variational mode decomposition (VMD) parameters. Then, the original passenger flow data containing nonlinear and nonstationary irregular changes of noise is decomposed into several intrinsic mode functions (IMFs) by using the optimized VMD technology, which reduces the time-varying complexity of passenger flow time series and improves predictability. Meanwhile, the IMFs are divided into different frequency series by fluctuation-based dispersion entropy, and diverse models are utilized to predict different frequency series. Finally, to avoid the cumulative error caused by the direct superposition of each IMF’s prediction result, reinforcement learning is adopted to ensemble the multiple models to acquire the multistep passenger flow prediction result. Experiments on four subway station passenger flow datasets proved that the prediction performance of the proposed method was better than all benchmark models. The excellent prediction effect of the proposed model has important guiding significance for evaluating the operation status of urban rail transit systems and improving the level of passenger service.
期刊介绍:
Urban Rail Transit is a peer-reviewed, international, interdisciplinary and open-access journal published under the SpringerOpen brand that provides a platform for scientists, researchers and engineers of urban rail transit to publish their original, significant articles on topics in urban rail transportation operation and management, design and planning, civil engineering, equipment and systems and other related topics to urban rail transit. It is to promote the academic discussions and technical exchanges among peers in the field. The journal also reports important news on the development and operating experience of urban rail transit and related government policies, laws, guidelines, and regulations. It could serve as an important reference for decision¬makers and technologists in urban rail research and construction field.
Specific topics cover:
Column I: Urban Rail Transportation Operation and Management
• urban rail transit flow theory, operation, planning, control and management
• traffic and transport safety
• traffic polices and economics
• urban rail management
• traffic information management
• urban rail scheduling
• train scheduling and management
• strategies of ticket price
• traffic information engineering & control
• intelligent transportation system (ITS) and information technology
• economics, finance, business & industry
• train operation, control
• transport Industries
• transportation engineering
Column II: Urban Rail Transportation Design and Planning
• urban rail planning
• pedestrian studies
• sustainable transport engineering
• rail electrification
• rail signaling and communication
• Intelligent & Automated Transport System Technology ?
• rolling stock design theory and structural reliability
• urban rail transit electrification and automation technologies
• transport Industries
• transportation engineering
Column III: Civil Engineering
• civil engineering technologies
• maintenance of rail infrastructure
• transportation infrastructure systems
• roads, bridges, tunnels, and underground engineering ?
• subgrade and pavement maintenance and performance
Column IV: Equipments and Systems
• mechanical-electronic technologies
• manufacturing engineering
• inspection for trains and rail
• vehicle-track coupling system dynamics, simulation and control
• superconductivity and levitation technology
• magnetic suspension and evacuated tube transport
• railway technology & engineering
• Railway Transport Industries
• transport & vehicle engineering
Column V: other topics of interest
• modern tram
• interdisciplinary transportation research
• environmental impacts such as vibration, noise and pollution
Article types:
• Papers. Reports of original research work.
• Design notes. Brief contributions on current design, development and application work; not normally more than 2500 words (3 journal pages), including descriptions of apparatus or techniques developed for a specific purpose, important experimental or theoretical points and novel technical solutions to commonly encountered problems.
• Rapid communications. Brief, urgent announcements of significant advances or preliminary accounts of new work, not more than 3500 words (4 journal pages). The most important criteria for acceptance of a rapid communication are novel and significant. For these articles authors must state briefly, in a covering letter, exactly why their works merit rapid publication.
• Review articles. These are intended to summarize accepted practice and report on recent progress in selected areas. Such articles are generally commissioned from experts in various field s by the Editorial Board, but others wishing to write a review article may submit an outline for preliminary consideration.