Reinforcement of QoE based Feedback System to Allocate Resources in IIoT
K. Udayakumar, S. Ramamoorthy · 2022 3rd International Conference for Emerging Technology (INCET) · 2022
Current IoT devices are demanding dynamic resource allocation policy to handle computation and communication delay. It also requires complex services of computing nodes to make functional decisions at the source end. Edge technology as a distributed paradigm allows execution closer to data source. Even though it allows local execution, a Resource Allocation(RA) scheme is required to ensure dynamic optimal allocation and utilization of computing resources. İn traditional RA scheme prefixed mapping table is used to allocate resources. It leads to computation delay in an IIoT environment where an optimal execution is feasible with reward-based models. Quality of Experience(QoE) based Dynamic RA Scheme for such a heterogeneous network could improve overall efficiency. In this paper, we considered Reinforcement Learning(RL) approach for resource allocation problem. The main objective is to propose a feedback-based dynamic RA model for IIoT using the RL approach. An agent in RL interacts with the environment to allocate resources and update allocation policy based on feedback. The experimental result shows that the proposed model converges to low cost.