Scholarlink

<p>The growing demand for services necessitates the deployment and integration of numerous fifth generation (5G) and beyond 5G networks to support extensive ultra-reliable low-latency communication (URLLC) services. This study aims to design an efficient network selection algorithm to fulfill the stringent requirements of high reliability and low latency for URLLC services by selecting the best network among multiple existing wireless access technologies throughout the service connection duration. Most of the existing network selection methods are less efficient and will not satisfy the quality of service (QoS) requirements. In order to meet the QoS in practical scenario involving a heterogeneous wireless network with uncertain environmental conditions, a model-free algorithm, specifically the Q-learning algorithm, is employed to address the problem within the framework of a markov decision process. The ultimate goal is to determine an optimal network that maximizes expected cumulative rewards. Finally, performance evaluation demonstrates when compare with Random Selection, Fuzzy Logic and multi attribute decision making methods such as the simple additive weighting and the Technique for Order Preference by Similarity to Ideal Solution, the proposed algorithm has higher expected total reward, reduced number of vertical handoff and efficient distribution of load among multiple networks.</p>

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

Q-learning based heterogeneous network selection decision algorithm for ultra reliable and low latency communication services

  • 期刊论文

  • 发表日期：

    2025年3月4日

查看全文
  • Abhishek Kumar Verma,
  • Vinay Kumar Singh,
  • M. R. Khan,
  • Prabira Kumar Sethy

摘要

The growing demand for services necessitates the deployment and integration of numerous fifth generation (5G) and beyond 5G networks to support extensive ultra-reliable low-latency communication (URLLC) services. This study aims to design an efficient network selection algorithm to fulfill the stringent requirements of high reliability and low latency for URLLC services by selecting the best network among multiple existing wireless access technologies throughout the service connection duration. Most of the existing network selection methods are less efficient and will not satisfy the quality of service (QoS) requirements. In order to meet the QoS in practical scenario involving a heterogeneous wireless network with uncertain environmental conditions, a model-free algorithm, specifically the Q-learning algorithm, is employed to address the problem within the framework of a markov decision process. The ultimate goal is to determine an optimal network that maximizes expected cumulative rewards. Finally, performance evaluation demonstrates when compare with Random Selection, Fuzzy Logic and multi attribute decision making methods such as the simple additive weighting and the Technique for Order Preference by Similarity to Ideal Solution, the proposed algorithm has higher expected total reward, reduced number of vertical handoff and efficient distribution of load among multiple networks.