{"title":"An α-vector predictive value iteration algorithm for transportation infrastructure maintenance under partially observable conditions","authors":"Chunhui Guo, Zhenglin Liang","doi":"10.1016/j.ress.2025.111235","DOIUrl":null,"url":null,"abstract":"<div><div>Transportation infrastructure is degrading over time and poses the risk of failure when exposed to a dynamic environment. Periodic inspection is often implemented to assess the requirement of maintenance. However, the inspected conditions can only partially reflect the underlying degradation, complicating the decision of maintenance. Moreover, inspections of early degradation often have no value-adding to condition improvement and incur a portion of unnecessary expenses. To address the abovementioned issues, we propose a sequential predictive maintenance policy that accounts for the partial observation of the infrastructure’s condition to reduce unnecessary inspections. The schedule of inspection timings is predicted according to the estimated Remaining Useful Life distribution, adaptive to stochastic degradation. We demonstrate that the optimal value function is piecewise linear and convex when decision epochs are non-periodic. Leveraging this insight, we have designed an <span><math><mi>α</mi></math></span>-vector Predictive Value Iteration algorithm (<span><math><mi>α</mi></math></span>-PVI) to optimize the transportation infrastructure maintenance policy. The <span><math><mi>α</mi></math></span>-PVI algorithm further reduces the time complexity compared with the Point-Based Value Iteration algorithm. Our designed approach is verified through an application for the maintenance optimization of pavements and bridges. The results demonstrate that the <span><math><mi>α</mi></math></span>-PVI algorithm reduces unnecessary inspection costs by on average 61.25% when compared to the periodic inspection approach. The <span><math><mi>α</mi></math></span>-PVI algorithm constructs a new paradigm of predictive maintenance under partially observable conditions.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"264 ","pages":"Article 111235"},"PeriodicalIF":11.0000,"publicationDate":"2025-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Reliability Engineering & System Safety","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0951832025004363","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, INDUSTRIAL","Score":null,"Total":0}
引用次数: 0
Abstract
Transportation infrastructure is degrading over time and poses the risk of failure when exposed to a dynamic environment. Periodic inspection is often implemented to assess the requirement of maintenance. However, the inspected conditions can only partially reflect the underlying degradation, complicating the decision of maintenance. Moreover, inspections of early degradation often have no value-adding to condition improvement and incur a portion of unnecessary expenses. To address the abovementioned issues, we propose a sequential predictive maintenance policy that accounts for the partial observation of the infrastructure’s condition to reduce unnecessary inspections. The schedule of inspection timings is predicted according to the estimated Remaining Useful Life distribution, adaptive to stochastic degradation. We demonstrate that the optimal value function is piecewise linear and convex when decision epochs are non-periodic. Leveraging this insight, we have designed an -vector Predictive Value Iteration algorithm (-PVI) to optimize the transportation infrastructure maintenance policy. The -PVI algorithm further reduces the time complexity compared with the Point-Based Value Iteration algorithm. Our designed approach is verified through an application for the maintenance optimization of pavements and bridges. The results demonstrate that the -PVI algorithm reduces unnecessary inspection costs by on average 61.25% when compared to the periodic inspection approach. The -PVI algorithm constructs a new paradigm of predictive maintenance under partially observable conditions.
期刊介绍:
Elsevier publishes Reliability Engineering & System Safety in association with the European Safety and Reliability Association and the Safety Engineering and Risk Analysis Division. The international journal is devoted to developing and applying methods to enhance the safety and reliability of complex technological systems, like nuclear power plants, chemical plants, hazardous waste facilities, space systems, offshore and maritime systems, transportation systems, constructed infrastructure, and manufacturing plants. The journal normally publishes only articles that involve the analysis of substantive problems related to the reliability of complex systems or present techniques and/or theoretical results that have a discernable relationship to the solution of such problems. An important aim is to balance academic material and practical applications.