Scholarlink

<p>This paper presents an anti-counterfeiting tracking and tracing method for power emergency materials based on a hypergraph neural network (HGNN) and a swarm intelligence algorithm. Each power emergency material is equipped with an RFID tag that stores multidimensional data, including production, transportation, and usage information. The RFID data are collected via NFC passive communication and fed into the HGNN model, which employs a heterogeneous hypergraph structure to capture high-order associations among tags. The model further integrates a vertex-edge attention aggregation network and a metapath aggregation network to enhance node feature extraction and semantic representation. As part of the swarm intelligence framework, an immune genetic algorithm (IGA) is used to optimize the hyperparameters of the HGNN. Experimental results show that the proposed method achieves an accuracy of 0.962, a recall of 0.948, an F1 score of 0.955, and a false positive rate of 0.033 on a simulated anti-counterfeiting classification test set. The Rand index of this method exceeds 0.95 in trajectory reduction for ten types of power emergency materials, and the solution space coverage rate surpasses 98% during the optimization process.</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

Anti-counterfeiting tracking and traceability method for power emergency supplies based on hypergraph neural network and swarm intelligence algorithm

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月31日

查看全文
  • Huayu Chu,
  • Lichong Cui,
  • Wei Guo,
  • Junsheng Wang,
  • Lei Su,
  • Si Wang

摘要

This paper presents an anti-counterfeiting tracking and tracing method for power emergency materials based on a hypergraph neural network (HGNN) and a swarm intelligence algorithm. Each power emergency material is equipped with an RFID tag that stores multidimensional data, including production, transportation, and usage information. The RFID data are collected via NFC passive communication and fed into the HGNN model, which employs a heterogeneous hypergraph structure to capture high-order associations among tags. The model further integrates a vertex-edge attention aggregation network and a metapath aggregation network to enhance node feature extraction and semantic representation. As part of the swarm intelligence framework, an immune genetic algorithm (IGA) is used to optimize the hyperparameters of the HGNN. Experimental results show that the proposed method achieves an accuracy of 0.962, a recall of 0.948, an F1 score of 0.955, and a false positive rate of 0.033 on a simulated anti-counterfeiting classification test set. The Rand index of this method exceeds 0.95 in trajectory reduction for ten types of power emergency materials, and the solution space coverage rate surpasses 98% during the optimization process.