预测性维护
停工期
可靠性(半导体)
预防性维护
强化学习
可靠性工程
过程(计算)
纠正性维护
工程类
主动维护
计算机科学
设备总体有效性
状态维修
生产力
维护措施
状态监测
制造业
最佳维护
自动引导车
自动化
实时计算
维修工程
启发式
作者
Maziyar Massahi,Hamid Tohidi,Seyedehfatemeh Golrizgashti
标识
DOI:10.1108/jqme-05-2025-0053
摘要
Purpose The purpose of this study is to enhance the maintenance performance of Automated Guided Vehicles (AGVs) in Industry 4.0 settings by developing an intelligent decision-making model. This model seeks to reduce corrective maintenance time and improve the reliability and productivity of AGVs through the implementation of optimized preventive and predictive maintenance strategies. Design/methodology/approach This research utilizes reinforcement learning to train a policy agent for optimizing maintenance decision-making for AGVs. The agent determines the optimal timing and type of maintenance actions, including preventive and predictive maintenance, with the aim of reducing corrective maintenance time and improving system reliability. In this study, the locations of the maintenance site and AGV charging station are predefined and examined through three alternative spatial scenarios to evaluate their effects on AGV performance. The model also incorporates buffer capacity constraints and combines simulation-based modeling with learning-based optimization to reflect real-world operational conditions in smart manufacturing environments. Findings The results demonstrate that the reinforcement learning agent can significantly reduce downtime and enhance AGV operational performance. Furthermore, the spatial configuration of maintenance site and AGV charging station has a measurable impact on system reliability and cost-efficiency. The integration of predictive maintenance strategies into the agent's learning process further stabilizes system operation and improves long-term equipment health. Research limitations/implications This study is limited to simulation-based validation and does not account for real-time uncertainties such as unexpected human intervention or sensor failures. Future research should focus on real-world implementation and consider multi-agent collaboration for fleet-level optimization. Practical implications The proposed framework can assist manufacturing firms in reducing maintenance costs and unplanned downtime while improving scheduling accuracy. It provides a decision-support tool for operations managers looking to implement intelligent maintenance within Industry 4.0 infrastructures. Originality/value This paper introduces a novel application of reinforcement learning to maintenance policy planning for AGVs, considering spatial and operational constraints. It offers value to both researchers and practitioners in the field of smart manufacturing, maintenance optimization, and AI-based industrial systems. Highlights
科研通智能强力驱动
Strongly Powered by AbleSci AI