Network Function Virtualisation (NFV) technology can flexibly manage choreographed network functions and provide a variety of network services, which has a positive effect on handling diverse IIoT service demands. Although NFV resource deployment has been extensively studied, there are still challenges in reducing the energy consumption of resource deployment while meeting the diverse IIoT service requirements. Therefore, this paper proposes EE-FDRL, an energy-efficient resource deployment algorithm based on Federated Reinforcement Learning, to reduce energy consumption and load balancing, for solving the resource deployment problem for diverse IIoT service requirements. To address the data privacy and security issues, the federated request-response method is used to achieve privacy protection by responding to shared request information instead of providing its own topology and resource information; to address the efficient deployment problem, the federated reinforcement learning algorithm based on the simultaneous deployment of node links is proposed to optimize the decision-making search space by integrating the deployment information into the reinforcement method to ensure that the service can be provided to meet the user’s to ensure the provision of services that meet the diversified needs of users under the resource constraints. To further improve the accuracy and efficiency of decision-making, the feature extraction and representation module is also configured so that the model can make full use of existing knowledge and experience. Simulation results show that the EE-FDRL method is superior to the proposed comparison algorithm.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Federated Reinforcement Learning-based Algorithm for Service Function Chain Deployment


    Beteiligte:
    Zhang, Yanli (Autor:in) / Yu, Tong (Autor:in) / Xu, Zexi (Autor:in) / Song, Yu (Autor:in)


    Erscheinungsdatum :

    23.10.2024


    Format / Umfang :

    1236643 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch






    Communication-efficient and federated multi-agent reinforcement learning

    Krouka, M. (Mounssif) / Elgabli, A. (Anis) / Issaid, C. B. (Chaouki Ben) et al. | BASE | 2022

    Freier Zugriff

    AVDDPG: Federated reinforcement learning applied to autonomous platoon control

    Boin, Christian / Lei, Lei / Yang, Simon X. | ArXiv | 2022

    Freier Zugriff