A reinforcement learning based approach for multi-projects scheduling in cloud manufacturing

Chen, Shengkai; Fang, Shuiliang; Tang, Renzhong

doi:10.1080/00207543.2018.1535205

Cited by 57 publications

(14 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…proposed a bio-inspired reference architecture for the cyberphysical production system to achieve the self-organization of the production system [48]. Chen et al proposed a reinforcement learning-based task-assigning policy to support multi-projects scheduling in the CMfg system [49].…”

Section: Autonomous Collaboration In Manufacturingmentioning

confidence: 99%

Social Production System: A Three-Layer Smart Framework for Implementing Autonomous Human-Machine Collaborations in a Shop Floor

2021

View full text Add to dashboard Cite

The continuous fast-changing of production requirements and the development of advanced information technology have promoted the development of the production system in a more knowledgeintensive, resources-decentralized, and autonomous collaborative manner. In order to provide guidelines for the smart transformation of the production shop floor, this paper proposes a smart production system with a three-layer framework defined as the social production system (Social-PS) by integrating cyber-physical system (CPS), knowledge graph technology, and production-event driven model. The concept of smart resources (include smart operator, smart machine, smart workpiece, and smart gateway), social sensor, the unified information model is clarified to construct the physical carrier of Social-PS. The unified knowledge repository's overall construction method based on the knowledge graph is elaborated to provide taskresource-associated heterogeneous production data sharing and management for human-machine collaboration. Subsequently, to support the production dynamic monitoring and interaction, a productionevent and active knowledge indexing model is proposed. Furthermore, to validate the proposed Social-PS framework, a software and hardware integrated prototype system is implemented to demonstrate the knowledge-driven human-machine collaboration of the production process, which provides a basis for realizing smart and autonomous production collaboration.

show abstract

Section: Autonomous Collaboration In Manufacturingmentioning

confidence: 99%

Social Production System: A Three-Layer Smart Framework for Implementing Autonomous Human-Machine Collaborations in a Shop Floor

2021

View full text Add to dashboard Cite

show abstract

“…However, it's difficult for RL methods (e.g., Q-learning) to take high-dimensional data as inputs and solve problems with large state-action spaces. Chen et al [30] used an RL-based assigning policy to obtain the non-dominated solution set in the action space that helps yield a better performance than Q-learning. Wang et al [31] used correlated equilibrium to propose a multi-agent RL algorithm for makespan and cost optimization to guide the scheduling of multi-workflows over clouds.…”

Section: Dynamic Scheduling Under Uncertaintymentioning

confidence: 99%

Reinforcement Learning With Composite Rewards for Production Scheduling in a Smart Factory

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Often the contributions present how to utilise an RL-agent to process specific operations waiting in the queue on a machine (Stricker et al 2018) or create complete production plans (Waschneck et al 2018). The application of allocation strategies in a stochastic environment combined with intensified learning can achieve up to 32% improvement in individual cases compared to conventional methods (Chen, Fang, and Tang 2019). Furthermore, an alternative machine can be selected in case of failure (Zhao et al 2019).…”

Section: Reinforcement Learningmentioning

confidence: 99%

Dynamically adjusting the k-values of the ATCS rule in a flexible flow shop scenario with reinforcement learning

Heger

Voß

2021

International Journal of Production Research

View full text Add to dashboard Cite

Given the fact that finding the optimal sequence in a flexible flow shop is usually an NP-hard problem, priority-based sequencing rules are applied in many real-world scenarios. In this contribution, an innovative reinforcement learning approach is used as a hyper-heuristic to dynamically adjust the k-values of the ATCS sequencing rule in a complex manufacturing scenario. For different product mixes as well as different utilisation levels, the reinforcement learning approach is trained and compared to the k-values found with an extensive simulation study. This contribution presents a human comprehensible hyper-heuristic, which is able to adjust the k-values to internal and external stimuli and can reduce the mean tardiness up to 5%.

show abstract

A reinforcement learning based approach for multi-projects scheduling in cloud manufacturing

Cited by 57 publications

References 38 publications

Social Production System: A Three-Layer Smart Framework for Implementing Autonomous Human-Machine Collaborations in a Shop Floor

Social Production System: A Three-Layer Smart Framework for Implementing Autonomous Human-Machine Collaborations in a Shop Floor

Reinforcement Learning With Composite Rewards for Production Scheduling in a Smart Factory

Dynamically adjusting the k-values of the ATCS rule in a flexible flow shop scenario with reinforcement learning

Contact Info

Product

Resources

About