Academic Journal
Multi-agent deep reinforcement learning-based maintenance optimization for multi-dependent component systems
| العنوان: | Multi-agent deep reinforcement learning-based maintenance optimization for multi-dependent component systems |
|---|---|
| المؤلفون: | Do, Phuc, Nguyen, Van-Thai, Voisin, Alexandre, Iung, Benoît, Neto, Waldomiro Alves Ferreira |
| المساهمون: | Centre de Recherche en Automatique de Nancy (CRAN), Université de Lorraine (UL)-Centre National de la Recherche Scientifique (CNRS), Universidade Federal de Pernambuco Recife (UFPE), European Project: 101091996,MODAPTO |
| المصدر: | ISSN: 0957-4174 ; Expert Systems with Applications ; https://hal.science/hal-04460407 ; Expert Systems with Applications, 2024, 245, pp.123144. ⟨10.1016/j.eswa.2024.123144⟩. |
| بيانات النشر: | HAL CCSD Elsevier |
| سنة النشر: | 2024 |
| المجموعة: | Université de Lorraine: HAL |
| مصطلحات موضوعية: | Deep reinforcement learning, Cooperative multi-agent systems, Maintenance decision-making, Multi-component systems, [SPI.AUTO]Engineering Sciences [physics]/Automatic |
| الوصف: | International audience ; Manufacturing systems consist of a set of interdependent components. However, addressing the dependence between these components remains a challenge in both maintenance modeling and the optimization process. In this paper, we propose a multi-agent deep reinforcement learning-based maintenance approach for a manufacturing system, taking into consideration both stochastic and economic dependencies between components. In this manner, we introduce a novel state interactions model, suggesting that the degradation state of one component may influence the degradation process of others. Subsequently, a maintenance planning approach based on multi-agent deep reinforcement learning is developed to optimize maintenance decisions in both fully and partially observed states. The deployed multi-agent deep reinforcement algorithm, specifically Weighted QMIX, ensures scalability and efficient consideration of state interactions and economic dependencies between components. The feasibility and performance of the proposed maintenance approach are investigated through various numerical studies. When compared to traditional maintenance approaches, such as value iteration method, Dueling Double Deep Q Network, and Multi-Agent Deep Q Network, our proposed approach consistently demonstrates superior results. |
| نوع الوثيقة: | article in journal/newspaper |
| اللغة: | English |
| Relation: | info:eu-repo/grantAgreement//101091996/EU/Manufacturing And Distributed Control Via Interoperable Digital Twins/MODAPTO; hal-04460407; https://hal.science/hal-04460407 |
| DOI: | 10.1016/j.eswa.2024.123144 |
| الاتاحة: | https://hal.science/hal-04460407 https://doi.org/10.1016/j.eswa.2024.123144 |
| رقم الانضمام: | edsbas.3926574E |
| قاعدة البيانات: | BASE |
| DOI: | 10.1016/j.eswa.2024.123144 |
|---|