Abstract:Problem definition: Unexpected component failures in industrial assets can lead to significant financial losses and operational disruptions, making preventive maintenance crucial for reducing unplanned downtime. This study focuses on optimizing maintenance for economically coupled assets that degrade according to a compound Poisson process, failing when degradation exceeds a specified threshold. Degradation parameters vary across components and cannot be directly observed, necessitating inference from real-time sensor data, with maintenance limited to scheduled interventions. Academic/practical relevance: We characterize optimal replacement policies that have access to the degradation process parameters, which inform suitable initial policies for our deep reinforcement learning (DRL) algorithm. The effectiveness of the trained policies is demonstrated through a case study on interventional X-ray systems. Methodology: The maintenance optimization problem is modeled as a partially observable Markov decision process (POMDP). To manage the inherent computational complexity, we reformulate it as a Bayesian Markov decision process (BMDP) using conjugate pairs. This allows us to extend the DRL algorithm to BMDPs. Additionally, we propose an open-loop feedback approach to adapt the DRL algorithm to the POMDP setting when a BMDP reformulation is infeasible. Results: Numerical results indicate that our DRL algorithm effectively reduces maintenance costs and unplanned downtime by leveraging real-time data and shared setup costs to learn state-of-the-art complex opportunistic maintenance policies. Managerial implications: Our approach demonstrates how integrating real-time learning and decision-making can significantly lower costs in industrial asset networks, providing a practical solution for improving operational efficiency amidst asset heterogeneity and economic dependence.

Predictive Maintenance Model for IIoT-based Manufacturing: A Transferable Deep Reinforcement Learning Approach

Deep-Reinforcement-Learning-Based Predictive Maintenance Model for Effective Resource Management in Industrial IoT

TranDRL: A Transformer-Driven Deep Reinforcement Learning Enabled Prescriptive Maintenance Framework

Deep-Learning-Enabled Predictive Maintenance in Industrial Internet of Things: Methods, Applications, and Challenges

Multi-agent deep reinforcement learning based Predictive Maintenance on parallel machines

Intelligent Maintenance Framework for Reconfigurable Manufacturing with Deep Learning-Based Prognostics

Deep reinforcement learning for maintenance optimization of a scrap-based steel production line

Real-Time Integrated Learning and Decision-Making for Asset Networks

Reinforcement learning for predictive maintenance: a systematic technical review

A digital twin to train deep reinforcement learning agent for smart manufacturing plants: Environment, interfaces and intelligence

Deep reinforcement learning in smart manufacturing: A review and prospects

IIoT-based Predictive Maintenance for Oil and Gas Industry

Deep-Reinforcement-Learning-Based Production Scheduling in Industrial Internet of Things

Probing an intelligent predictive maintenance approach with deep learning and augmented reality for machine tools in IoT-enabled manufacturing

Improving Cognitive Ability of Edge Intelligent IIoT Through Machine Learning

Intelligent Fault Quantitative Identification for Industrial Internet of Things (IIoT) via a Novel Deep Dual Reinforcement Learning Model Accompanied With Insufficient Samples

An Augmented Reality-Assisted Prognostics and Health Management System Based on Deep Learning for IoT-Enabled Manufacturing

Deep Reinforcement Learning Assisted Federated Learning Algorithm for Data Management of IIoT

Research on Deep Reinforcement Learning Based Intelligent Shop Scheduling Method

Deep Reinforcement Learning Based Intelligent Job Batching in Industrial Internet of Things

Deep learning models for predictive maintenance: a survey, comparison, challenges and prospects