Scholarlink

Now a days, due to the nonlinear increase of IoT devices in the digital world, massive amount of data is generated per second. To process and to complete the task submitted by User Equipment’s (UE’s) at lower level layer an efficient Cloud-Fog environment is needed. The task submitted by the UE’s and node configuration in cloud environment are heterogeneous in nature. So, fitting the task in the optimal cloud node is a challenging task. Moreover, today the UE’s tasks are demanding Low latency, Energy consumption and enhanced user QoE parameters etc., which could be attained with Cloud-Fog environment. In this paper we propose an efficient Reinforcement Learning based Resource allocation in Cloud-Fog environment (RL-Realoc). In the first phase the higher configuration nodes in the target environment are identified using K-means clustering. Then, the identified nodes cluster and UE’s requirements are analyzed for optimal node selection using Reinforcement Learning. The proposed system (RL-Realoc) ensures less execution cost, energy consumption hence reduces the fronthaul traffic with enhanced QoE parameters.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Reinforcement Learning Based Heterogeneous Resource Management in Cloud – Fog Environment

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • R. S. Vindan,
  • M. Gobi,
  • Karthik Mohan,
  • T. Suriya Praba,
  • V. Meena

摘要

Now a days, due to the nonlinear increase of IoT devices in the digital world, massive amount of data is generated per second. To process and to complete the task submitted by User Equipment’s (UE’s) at lower level layer an efficient Cloud-Fog environment is needed. The task submitted by the UE’s and node configuration in cloud environment are heterogeneous in nature. So, fitting the task in the optimal cloud node is a challenging task. Moreover, today the UE’s tasks are demanding Low latency, Energy consumption and enhanced user QoE parameters etc., which could be attained with Cloud-Fog environment. In this paper we propose an efficient Reinforcement Learning based Resource allocation in Cloud-Fog environment (RL-Realoc). In the first phase the higher configuration nodes in the target environment are identified using K-means clustering. Then, the identified nodes cluster and UE’s requirements are analyzed for optimal node selection using Reinforcement Learning. The proposed system (RL-Realoc) ensures less execution cost, energy consumption hence reduces the fronthaul traffic with enhanced QoE parameters.