Scholarlink

<p>Resource allocation in the Internet of Things (IoT) is challenging due to the dynamic and heterogeneous nature of devices. Traditional optimization methods like Simulated Annealing (SA) face limitations in adapting to real-time conditions. This paper presents a hybrid approach combining SA with Modified Deep Reinforcement Learning (MDRL) to address these challenges, enhancing energy efficiency, reducing latency, and improving scalability in dynamic IoT environments. SA is used to perform a global search for optimal resource allocation, while MDRL dynamically adjusts the allocation policies based on real-time feedback, enhancing adaptability. This carefully engineered hybrid strategy leverages the strengths of both optimization techniques, offering improvements in energy efficiency, network delay, and scalability, as shown through extensive simulations. Although the integration of SA with MDRL offers a novel approach in the context of IoT networks, it represents an incremental advancement in optimization techniques rather than a fundamentally new paradigm. The results indicate that the proposed strategy outperforms traditional methods, demonstrating practical advantages for large-scale, dynamic IoT environments.</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

Smart Resource Management in IoT: A Hybrid Approach with Simulated Annealing and Deep Reinforcement Learning

  • 期刊论文

  • 发表日期：

    2026年2月16日

查看全文
  • G. Kalingarani,
  • P. Selvaraj

摘要

Resource allocation in the Internet of Things (IoT) is challenging due to the dynamic and heterogeneous nature of devices. Traditional optimization methods like Simulated Annealing (SA) face limitations in adapting to real-time conditions. This paper presents a hybrid approach combining SA with Modified Deep Reinforcement Learning (MDRL) to address these challenges, enhancing energy efficiency, reducing latency, and improving scalability in dynamic IoT environments. SA is used to perform a global search for optimal resource allocation, while MDRL dynamically adjusts the allocation policies based on real-time feedback, enhancing adaptability. This carefully engineered hybrid strategy leverages the strengths of both optimization techniques, offering improvements in energy efficiency, network delay, and scalability, as shown through extensive simulations. Although the integration of SA with MDRL offers a novel approach in the context of IoT networks, it represents an incremental advancement in optimization techniques rather than a fundamentally new paradigm. The results indicate that the proposed strategy outperforms traditional methods, demonstrating practical advantages for large-scale, dynamic IoT environments.