Due to the increase in resource-constrained internet of things (IoT) devices, the multi-access edge computing (MEC) have become very competitive environments in terms of successful data offloading and allocation of computational resources. This competition together with the varying workload and service requests makes the real deployment of computational resources in edge servers a major challenge. In order to address this challenge, an intelligent offloading and container provisioning scheme is developed using a prospect-theoretic deep reinforcement learning (DRL) strategy. An offloading utility function is formulated by exploiting offloading overhead options, and a neural network (NN) is used to monitor queue states and workload matching to construct a behavioral function that will influence accurate container provisioning. A scaled cost function is formulated to balance the energy consumption and the quality of service (QoS) cost subject to hard per-task latency constraints, which lead to better container utilization and energy consumption.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Prospect-theoretic DRL Approach for Container Provisioning in Energy-constrained Edge Platforms


    Contributors:


    Publication date :

    2023-06-01


    Size :

    1228914 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English