Scholarlink

<p>The efficiency of opportunistic Q-learning-based routing protocols in mobile ad hoc networks depends on how well the agent’s exploration—exploitation dilemma is managed. This paper focuses on enhancing the decision-making process of a Q-learning-based reinforcement learning agent by optimizing the Candidate Forwarder List (CFL) in an opportunistic routing protocol for Mobile Ad Hoc Networks (MANETs). A single-cluster Density-Based Spatial Clustering of Applications with Noise (DBSCAN) technique is used to generate the First Forwarder List (FFL), which includes nodes with residual energy above 25% and located within 250 meters of the current node. To form the CFL, a mobility-aware threshold distance is computed using the relative speed and distance of neighboring nodes. Nodes that meet this threshold are considered stable and are added to the CFL, reducing link breakage. A modified epsilon-greedy strategy is applied, where the exploration decision is influenced by the mobility threshold, ensuring that only mobility-aware and stable nodes are selected. This balances exploration and exploitation during learning. Simulation results demonstrate that the proposed method achieves a Packet Delivery Ratio (PDR) of 87% and an average end-to-end delay of 0.06 seconds. The exploration-to-exploitation ratio is around 1.033, confirming the effectiveness of the approach in highly dynamic and energy-constrained MANET 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

Energy aware DBSCAN and mobility aware balanced q-learning based opportunistic routing protocol in MANET

  • 期刊论文

  • 发表日期：

    2025年5月16日

查看全文
  • A. Daniel Gurwin Das,
  • E. Anna Devi

摘要

The efficiency of opportunistic Q-learning-based routing protocols in mobile ad hoc networks depends on how well the agent’s exploration—exploitation dilemma is managed. This paper focuses on enhancing the decision-making process of a Q-learning-based reinforcement learning agent by optimizing the Candidate Forwarder List (CFL) in an opportunistic routing protocol for Mobile Ad Hoc Networks (MANETs). A single-cluster Density-Based Spatial Clustering of Applications with Noise (DBSCAN) technique is used to generate the First Forwarder List (FFL), which includes nodes with residual energy above 25% and located within 250 meters of the current node. To form the CFL, a mobility-aware threshold distance is computed using the relative speed and distance of neighboring nodes. Nodes that meet this threshold are considered stable and are added to the CFL, reducing link breakage. A modified epsilon-greedy strategy is applied, where the exploration decision is influenced by the mobility threshold, ensuring that only mobility-aware and stable nodes are selected. This balances exploration and exploitation during learning. Simulation results demonstrate that the proposed method achieves a Packet Delivery Ratio (PDR) of 87% and an average end-to-end delay of 0.06 seconds. The exploration-to-exploitation ratio is around 1.033, confirming the effectiveness of the approach in highly dynamic and energy-constrained MANET environments.