Decision at First Sight: An Attention Network for Direct Maintenance Optimization from Sensor Data
##plugins.themes.bootstrap3.article.main##
##plugins.themes.bootstrap3.article.sidebar##
Abstract
Maintenance planning is a crucial strategy in industrial systems, where maintenance costs can consume up to 40% of total production expenses, and downtime costs can reach hundreds of thousands of dollars per hour. Despite its importance, the implementation of advanced maintenance approaches remains limited due to challenges such as insufficient resources, lack of expertise, inadequate funding, and difficulty converting vast operational data into actionable decisions. This paper introduces a novel attention-based deep learning model for maintenance scheduling that bypasses traditional degradation modeling and optimization techniques. The proposed model operates directly on sensor data, leveraging a multi-head attention mechanism within an encoder-decoder architecture to generate maintenance schedules. The cost function of the model is flexible and can be customized to accommodate different maintenance scenarios, making it adaptable to various operational requirements. The model's performance is validated through comparisons with the state-of-the-art predict-then-optimize benchmark, demonstrating its ability to generate cost-effective maintenance schedules.
For commercial lithium-ion battery fleets, ATOM achieves a 22–35% reduction in maintenance expenses relative to predict-then-optimize approach.
This approach provides a scalable, data-driven solution for dynamic and complex maintenance environments, eliminating the need for explicit remaining useful life (RUL) estimates or predefined degradation models.
How to Cite
##plugins.themes.bootstrap3.article.details##
multi-head attention networks, foundation models, end-to-end optimization, condition-based maintenance, predictive maintenance
Ali, A., & Abdelhadi, A. (2022). Condition-based monitoring and maintenance: State of the art review. Applied Sciences, 12(2), 688.
Bastani, H., Zhang, D. J., & Zhang, H. (2022). Applied machine learning in operations management. Innovative Technology at the Interface of Finance and Operations: Volume I, 189–222.
Carvalho, T. P., Soares, F. A., Vita, R., Francisco, R. d. P., Basto, J. P., & Alcalá, S. G. (2019). A systematic literature review of machine learning methods applied to predictive maintenance. Computers & Industrial Engineering, 137, 106024.
Deloitte. (2017). Predictive maintenance and the smart factory. Deloitte Insights. https://www2.deloitte.com/us/en/insights/focus/industry-4-0/predictive-maintenance-smart-factory.html
Deloitte, & Institute, T. M. (2018). 2018 skills gap in manufacturing study. Deloitte Insights. https://www2.deloitte.com/us/en/pages/manufacturing/articles/future-of-manufacturing-skills-gap-study.html
Elbasheer, M., Longo, F., Mirabelli, G., Padovano, A., Solina, V., & Talarico, S. (2022). Integrated prescriptive maintenance and production planning: A machine learning approach for the development of an autonomous decision support agent. IFAC-PapersOnLine, 55(10), 2605–2610.
Elmachtoub, A. N., & Grigas, P. (2022). Smart “predict, then optimize”. Management Science, 68(1), 9–26.
Engineering, P. (2018). 2018 maintenance study. Plant Engineering Magazine. https://www.plantengineering.com/articles/2018-maintenance-report/
Fallahi, F., Bakir, I., Yildirim, M., & Ye, Z. (2022). A chance-constrained optimization framework for wind farms to manage fleet-level availability in condition-based maintenance and operations. Renewable and Sustainable Energy Reviews, 168, 112789.
Lee, J., & Mitici, M. (2023). Deep reinforcement learning for predictive aircraft maintenance using probabilistic remaining-useful-life prognostics. Reliability Engineering & System Safety, 230, 108908.
Li, Y., Peng, S., Li, Y., & Jiang, W. (2020). A review of condition-based maintenance: Its prognostic and operational aspects. Frontiers of Engineering Management, 7(3), 323–334.
Lv, Y., Guo, X., Zhou, Q., Qian, L., & Liu, J. (2023). Predictive maintenance decision-making for variable faults with non-equivalent costs of fault severities. Advanced Engineering Informatics, 56, 102011.
Manyika, J. (2011). Big data: The next frontier for innovation, competition, and productivity. McKinsey Global Institute, 1.
PwC, & Mainnovation. (2017). Predictive maintenance 4.0: Beyond the hype. https://www.pwc.nl/en/publicaties/predictive-maintenance-4-0.html
Quatrini, E., Costantino, F., Di Gravio, G., & Patriarca, R. (2020). Condition-based maintenance—an extensive literature review. Machines, 8(2), 31.
Research, A. (2016). The critical cost of downtime. https://www.aberdeen.com/opspro-essentials/critical-cost-downtime/
Rezaeianjouybari, B., & Shang, Y. (2020). Deep learning for prognostics and health management: State of the art, challenges, and opportunities. Measurement, 163, 107929.
Rupprecht, T., & Wang, Y. (2022). A survey for deep reinforcement learning in Markovian cyber–physical systems: Common problems and solutions. Neural Networks, 153, 13–36.
Sadana, U., Chenreddy, A., Delage, E., Forel, A., Frejinger, E., & Vidal, T. (2024). A survey of contextual optimization methods for decision-making under uncertainty. European Journal of Operational Research.
Saeed, A., A. Khan, M., Akram, U., J. Obidallah, W., Jawed, S., & Ahmad, A. (2025). Deep learning-based approaches for intelligent industrial machinery health management and fault diagnosis in resource-constrained environments. Scientific Reports, 15(1), 1114.
Sharma, J., Mittal, M. L., & Soni, G. (2024). Condition-based maintenance using machine learning and role of interpretability: A review. International Journal of System Assurance Engineering and Management, 15(4), 1345–1360.
Wang, W. (2012). An overview of the recent advances in delay-time-based maintenance modelling. Reliability Engineering & System Safety, 106, 165–178.
Wang, Y., Zhao, Y., & Addepalli, S. (2020). Remaining useful life prediction using deep learning approaches: A review. Procedia Manufacturing, 49, 81–88.
Wesendrup, K., & Hellingrath, B. (2023). Post-prognostics demand management, production, spare parts and maintenance planning for a single-machine system using reinforcement learning. Computers & Industrial Engineering, 179, 109216.
Williams, R. J. (1992). Simple statistical gradient-following algorithms for connectionist reinforcement learning. Machine Learning, 8, 229–256.
Xiang, X., & Foo, S. (2021). Recent advances in deep reinforcement learning applications for solving partially observable Markov decision processes (POMDP) problems: Part 1—fundamentals and applications in games, robotics and natural language processing. Machine Learning and Knowledge Extraction, 3(3), 554–581.
Yildirim, M., Gebraeel, N. Z., & Sun, X. A. (2017). Integrated predictive analytics and optimization for opportunistic maintenance and operations in wind farms. IEEE Transactions on Power Systems, 32(6), 4319–4328.
Yildirim, M., Sun, X. A., & Gebraeel, N. Z. (2016). Sensor-driven condition-based generator maintenance scheduling—Part I: Maintenance problem. IEEE Transactions on Power Systems, 31(6), 4253–4262.
Zhang, Y., Zhang, D., Zhang, X., Qiu, L., Chan, F. T., Wang, Z., & Zhang, S. (2023). Guided probabilistic reinforcement learning for sampling-efficient maintenance scheduling of multi-component system. Applied Mathematical Modelling, 119, 677–697.
Zhu, J., Wang, Y., Huang, Y., Bhushan Gopaluni, R., Cao, Y., Heere, M., . . . others. (2022). Data-driven capacity estimation of commercial lithium-ion batteries from voltage relaxation. Nature Communications, 13(1), 2261.
This work is licensed under a Creative Commons Attribution 3.0 Unported License.
The Prognostic and Health Management Society advocates open-access to scientific data and uses a Creative Commons license for publishing and distributing any papers. A Creative Commons license does not relinquish the author’s copyright; rather it allows them to share some of their rights with any member of the public under certain conditions whilst enjoying full legal protection. By submitting an article to the International Conference of the Prognostics and Health Management Society, the authors agree to be bound by the associated terms and conditions including the following:
As the author, you retain the copyright to your Work. By submitting your Work, you are granting anybody the right to copy, distribute and transmit your Work and to adapt your Work with proper attribution under the terms of the Creative Commons Attribution 3.0 United States license. You assign rights to the Prognostics and Health Management Society to publish and disseminate your Work through electronic and print media if it is accepted for publication. A license note citing the Creative Commons Attribution 3.0 United States License as shown below needs to be placed in the footnote on the first page of the article.
First Author et al. This is an open-access article distributed under the terms of the Creative Commons Attribution 3.0 United States License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.