Deep reinforcement learning for dynamic scheduling of a flexible job shop
The ability to handle unpredictable dynamic events is becoming more important in pursuing agile and flexible production scheduling. At the same time, the cyber-physical convergence in production system creates massive amounts of industrial data that needs to be mined and analysed in real-time. To fa...
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Journal Article |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/163903 |
| _version_ | 1826115662435581952 |
|---|---|
| author | Liu, Renke Piplani, Rajesh Toro, Carlos |
| author2 | School of Mechanical and Aerospace Engineering |
| author_facet | School of Mechanical and Aerospace Engineering Liu, Renke Piplani, Rajesh Toro, Carlos |
| author_sort | Liu, Renke |
| collection | NTU |
| description | The ability to handle unpredictable dynamic events is becoming more important in pursuing agile and flexible production scheduling. At the same time, the cyber-physical convergence in production system creates massive amounts of industrial data that needs to be mined and analysed in real-time. To facilitate such real-time control, this research proposes a hierarchical and distributed architecture to solve the dynamic flexible job shop scheduling problem. Double Deep Q-Network algorithm is used to train the scheduling agents, to capture the relationship between production information and scheduling objectives, and make real-time scheduling decisions for a flexible job shop with constant job arrivals. Specialised state and action representations are proposed to handle the variable specification of the problem in dynamic scheduling. Additionally, a surrogate reward-shaping technique to improve learning efficiency and scheduling effectiveness is developed. A simulation study is carried out to validate the performance of the proposed approach under different scenarios. Numerical results show that not only does the proposed approach deliver superior performance as compared to existing scheduling strategies, its advantages persist even if the manufacturing system configuration changes. |
| first_indexed | 2024-10-01T03:58:48Z |
| format | Journal Article |
| id | ntu-10356/163903 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2024-10-01T03:58:48Z |
| publishDate | 2022 |
| record_format | dspace |
| spelling | ntu-10356/1639032022-12-21T07:19:35Z Deep reinforcement learning for dynamic scheduling of a flexible job shop Liu, Renke Piplani, Rajesh Toro, Carlos School of Mechanical and Aerospace Engineering Engineering::Industrial engineering Dynamic Scheduling Flexible Job Shop The ability to handle unpredictable dynamic events is becoming more important in pursuing agile and flexible production scheduling. At the same time, the cyber-physical convergence in production system creates massive amounts of industrial data that needs to be mined and analysed in real-time. To facilitate such real-time control, this research proposes a hierarchical and distributed architecture to solve the dynamic flexible job shop scheduling problem. Double Deep Q-Network algorithm is used to train the scheduling agents, to capture the relationship between production information and scheduling objectives, and make real-time scheduling decisions for a flexible job shop with constant job arrivals. Specialised state and action representations are proposed to handle the variable specification of the problem in dynamic scheduling. Additionally, a surrogate reward-shaping technique to improve learning efficiency and scheduling effectiveness is developed. A simulation study is carried out to validate the performance of the proposed approach under different scenarios. Numerical results show that not only does the proposed approach deliver superior performance as compared to existing scheduling strategies, its advantages persist even if the manufacturing system configuration changes. 2022-12-21T07:19:35Z 2022-12-21T07:19:35Z 2022 Journal Article Liu, R., Piplani, R. & Toro, C. (2022). Deep reinforcement learning for dynamic scheduling of a flexible job shop. International Journal of Production Research, 60(13), 4049-4069. https://dx.doi.org/10.1080/00207543.2022.2058432 0020-7543 https://hdl.handle.net/10356/163903 10.1080/00207543.2022.2058432 2-s2.0-85129220142 13 60 4049 4069 en International Journal of Production Research © 2022 Informa UK Limited, trading as Taylor & Francis Group. All rights reserved. |
| spellingShingle | Engineering::Industrial engineering Dynamic Scheduling Flexible Job Shop Liu, Renke Piplani, Rajesh Toro, Carlos Deep reinforcement learning for dynamic scheduling of a flexible job shop |
| title | Deep reinforcement learning for dynamic scheduling of a flexible job shop |
| title_full | Deep reinforcement learning for dynamic scheduling of a flexible job shop |
| title_fullStr | Deep reinforcement learning for dynamic scheduling of a flexible job shop |
| title_full_unstemmed | Deep reinforcement learning for dynamic scheduling of a flexible job shop |
| title_short | Deep reinforcement learning for dynamic scheduling of a flexible job shop |
| title_sort | deep reinforcement learning for dynamic scheduling of a flexible job shop |
| topic | Engineering::Industrial engineering Dynamic Scheduling Flexible Job Shop |
| url | https://hdl.handle.net/10356/163903 |
| work_keys_str_mv | AT liurenke deepreinforcementlearningfordynamicschedulingofaflexiblejobshop AT piplanirajesh deepreinforcementlearningfordynamicschedulingofaflexiblejobshop AT torocarlos deepreinforcementlearningfordynamicschedulingofaflexiblejobshop |