Marchesano, Maria Grazia (2023) Advancement in Operations Management through Deep Reinforcement Learning and Simulation. [Tesi di dottorato]
Anteprima |
Testo
Marchesano_Maria Grazia_36.pdf Download (9MB) | Anteprima |
| Tipologia del documento: | Tesi di dottorato |
|---|---|
| Lingua: | English |
| Titolo: | Advancement in Operations Management through Deep Reinforcement Learning and Simulation |
| Autori: | Autore Email Marchesano, Maria Grazia mariagrazia.marchesano@unina.it |
| Data: | 13 Dicembre 2023 |
| Numero di pagine: | 143 |
| Istituzione: | Università degli Studi di Napoli Federico II |
| Dipartimento: | Ingegneria Chimica, dei Materiali e della Produzione Industriale |
| Dottorato: | Ingegneria dei prodotti e dei processi industriali |
| Ciclo di dottorato: | 36 |
| Coordinatore del Corso di dottorato: | nome email D'Anna, Andrea anddanna@unina.it |
| Tutor: | nome email Guizzi, Guido [non definito] |
| Data: | 13 Dicembre 2023 |
| Numero di pagine: | 143 |
| Parole chiave: | Operations Management, Deep Reinforcement Learning (DRL), Flow Shop, Maintenance Scheduling, Battery Swapping Station, Smart Microgrid. |
| Settori scientifico-disciplinari del MIUR: | Area 09 - Ingegneria industriale e dell'informazione > ING-IND/17 - Impianti industriali meccanici |
| Depositato il: | 08 Gen 2024 10:35 |
| Ultima modifica: | 12 Ago 2026 05:36 |
| URI: | https://www.fedoa.unina.it/id/eprint/15622 |
Abstract
The profound impact of artificial intelligence and advanced simulations has transformed various aspects of operations management, addressing complexities and optimizing processes in contemporary manufacturing and energy landscapes. This comprehensive research delves into three primary domains: production control, maintenance scheduling, and the integration of Battery Swapping Station (BSS) into smart microgrids. In the contest of production control, the study highlights the imperative nature of effectively coordinating material flow, resources, and processes, with a keen focus on Work in Process (WIP) control. A groundbreaking technique has been introduced that integrates Reinforcement Learning and Industry 4.0 technologies, aiming to regulate throughput in a Flow Shop setting and dynamically define sequencing rules. By employing a deep Q-network (DQN), the study achieves a targeted throughput while maintaining constant WIP. A central aspect of this work is the adaptability of the proposed "intelligent" tool, which adjusts its decision-making contingent on the dynamic changes in a production environment. Shifting focus to maintenance scheduling in Flow Shop production lines, Deep Reinforcement Learning (DRL) emerges as a game-changer. The research proposes an innovative integration of DRL and simulation tools to optimize maintenance tasks scheduling, harnessing the predictive power of simulations and the adaptive capabilities of DRL. This approach's uniqueness lies in training on a single machine and testing the resulting policy across varying machine configurations, demonstrating its robustness in diverse experimental settings. Lastly, the research addresses the potential of integrating Battery Swapping Station (BSS) within a smart microgrid. A detailed simulation model, acting as a digital twin of the microgrid, is developed, capturing the nuances of renewable energy sources, second-life battery storage, and utilities. The study meticulously evaluates the economic viability of this integration, emphasizing cost optimization and effective energy demand management. Notably, the initiative promotes the circular economy by championing the utilization of second-life batteries. In conclusion, this research accentuates the transformative potential of AI-driven techniques and advanced simulations in redefining operations management paradigms across manufacturing and energy sectors.
Downloads
Downloads per month over past year
Actions (login required)
 |
Modifica documento |
