Deep reinforcement learning in smart manufacturing: A review and prospects

CIRP journal of manufacturing science and technology, Journal Year: 2022, Volume and Issue: 40, P. 75 - 101

Published: Dec. 2, 2022

Language: Английский

---

Dynamic job shop scheduling based on deep reinforcement learning for multi-agent manufacturing systems

Robotics and Computer-Integrated Manufacturing, Journal Year: 2022, Volume and Issue: 78, P. 102412 - 102412

Published: July 6, 2022

Language: Английский

---

Flexible job shop scheduling problem under Industry 5.0: A survey on human reintegration, environmental consideration and resilience improvement

Journal of Manufacturing Systems, Journal Year: 2023, Volume and Issue: 67, P. 155 - 173

Published: Jan. 28, 2023

Language: Английский

---

Solving job scheduling problems in a resource preemption environment with multi-agent reinforcement learning

Robotics and Computer-Integrated Manufacturing, Journal Year: 2022, Volume and Issue: 77, P. 102324 - 102324

Published: March 3, 2022

Language: Английский

---

Survey on Genetic Programming and Machine Learning Techniques for Heuristic Design in Job Shop Scheduling

IEEE Transactions on Evolutionary Computation, Journal Year: 2023, Volume and Issue: 28(1), P. 147 - 167

Published: March 10, 2023

Job shop scheduling (JSS) is a process of optimizing the use limited resources to improve production efficiency. JSS has wide range applications, such as order picking in warehouse and vaccine delivery under pandemic. In real-world environment often complex due dynamic events, job arrivals over time machine breakdown. Scheduling heuristics, e.g., dispatching rules, have been popularly used prioritize candidates machines manufacturing make good schedules efficiently. Genetic programming (GP), shown its superiority learning heuristics for automatically flexible representation. This survey first provides comprehensive discussions recent designs GP algorithms on different types JSS. addition, we notice that years, techniques, feature selection multitask learning, adapted effectiveness efficiency heuristic design with GP. However, there no discuss strengths weaknesses these approaches. To fill this gap, article techniques automatic current issues challenges are discussed identify promising areas future.

Language: Английский

---

A systematic review and meta-analysis of machine learning, deep learning, and ensemble learning approaches in predicting EV charging behavior

Engineering Applications of Artificial Intelligence, Journal Year: 2024, Volume and Issue: 135, P. 108789 - 108789

Published: July 4, 2024

Language: Английский

---

A dual population collaborative genetic algorithm for solving flexible job shop scheduling problem with AGV

Swarm and Evolutionary Computation, Journal Year: 2024, Volume and Issue: 86, P. 101538 - 101538

Published: March 11, 2024

Language: Английский

---

A Learning-Driven Multi-Objective cooperative artificial bee colony algorithm for distributed flexible job shop scheduling problems with preventive maintenance and transportation operations

Computers & Industrial Engineering, Journal Year: 2024, Volume and Issue: 196, P. 110484 - 110484

Published: Aug. 18, 2024

Language: Английский

---

Deep reinforcement learning-based memetic algorithm for energy-aware flexible job shop scheduling with multi-AGV

Computers & Industrial Engineering, Journal Year: 2024, Volume and Issue: 189, P. 109917 - 109917

Published: Feb. 2, 2024

Language: Английский

---

Job shop smart manufacturing scheduling by deep reinforcement learning

Journal of Industrial Information Integration, Journal Year: 2024, Volume and Issue: 38, P. 100582 - 100582

Published: Feb. 9, 2024

Smart manufacturing scheduling (SMS) requires a high degree of flexibility to successfully cope with changes in operational decision level planning processes today's production environments, which are usually subject uncertainty. In such unique and complex scenario as the real job shop, modelling SMS Markov process (MDP), its approach by deep reinforcement learning (DRL), is research field growing interest given characteristics. It allows us consider achieving levels promoting automation, autonomy making, ability act time when faced disturbances disruptions highly dynamic environment. This paper addresses problem quasi-realistic shop environment characterised machines receiving jobs from buffers that accumulate numerous using wide variety parts multimachine routes diverse number operation phases developing digital twin based on MDP DRL methodology. approached by: OpenAI Gym; designing an observation space 18 features; action composed three priority heuristic rules; shaping single reward function multi-objective characteristic; implementation proximal policy optimisation (PPO) algorithm Stable Baselines 3 library. approach, dubbed smart (JS-SMS), deterministic formulation implementation. The model subjected validation comparing it several best-known rules. main findings this methodology allow replicate, great extent, positive aspects rules mitigate negative ones, achieves more balanced behaviour most measures established performance indicators outperforms perspective. Finally, further oriented stochastic approaches address reality Industry 4.0 context.

Language: Английский

---