Scholarlink

<p>Supply chain management (SCM) actively contributes to the supply chain and logistic including coordination and integration of supply and demand management within Indian manufacturing industries. The rise of IoT (Internet of Things) has great significance in SCM process making it more smart and efficient for manufacturing sector. This study presents a real-time SCM analysis using IoT-enabled data that is collected across multiple logistics stages. The proposed MDHO-GNN-GRU computational framework is composed of Monkey Drill Hybrid Optimization (MDHO) for selecting optimized feature from full feature space, Graph Neural Networks (GNN) for modeling complex relationships, and GRU (Gated recurrent unit) component with attention mechanism captures temporal dependencies. This MDHO metaheuristic approach is also efficient for tuning hyperparameters to achieve optimization. The fusion of combining outputs form GNN and GRU is a key point that enables deep dynamic system for relational structures among IoT devices and capturing temporal patterns. GNN significantly captures interdependencies in dynamic supply networks and multi-agent systems. GRU enables efficient temporal learning with fewer parameters than LSTMs or Transformers. Performance scores show an accuracy of 99.7% and precision of 98.3% thus outperforming recent approaches. Thus, the proposed approach offers valuable insights for proactive SCM in manufacturing sector.</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

MDHO-GNN-GRUNet: a hybrid IoT and graph-gated recurrent framework for sustainable supply chain performance in Indian manufacturing industries

  • 期刊论文

  • 发表日期：

    2025年11月21日

查看全文
  • Arpit Singh,
  • Sachin Saini

摘要

Supply chain management (SCM) actively contributes to the supply chain and logistic including coordination and integration of supply and demand management within Indian manufacturing industries. The rise of IoT (Internet of Things) has great significance in SCM process making it more smart and efficient for manufacturing sector. This study presents a real-time SCM analysis using IoT-enabled data that is collected across multiple logistics stages. The proposed MDHO-GNN-GRU computational framework is composed of Monkey Drill Hybrid Optimization (MDHO) for selecting optimized feature from full feature space, Graph Neural Networks (GNN) for modeling complex relationships, and GRU (Gated recurrent unit) component with attention mechanism captures temporal dependencies. This MDHO metaheuristic approach is also efficient for tuning hyperparameters to achieve optimization. The fusion of combining outputs form GNN and GRU is a key point that enables deep dynamic system for relational structures among IoT devices and capturing temporal patterns. GNN significantly captures interdependencies in dynamic supply networks and multi-agent systems. GRU enables efficient temporal learning with fewer parameters than LSTMs or Transformers. Performance scores show an accuracy of 99.7% and precision of 98.3% thus outperforming recent approaches. Thus, the proposed approach offers valuable insights for proactive SCM in manufacturing sector.