A Scalable and Energy-Efficient Clustering Framework for IIoT-Enabled WSNs Using Enhanced Grey Wolf Optimizer and GraphSAGE-Based Learning

In Industrial Internet of Things (IIoT) environments and WSNs effective clustering and optimal Cluster Head (CH) selection are essential for prolonging network lifespan, minimizing energy consumption, and maintaining robust connectivity. This study introduces an advanced clustering framework that synergistically combines an Enhanced Grey Wolf Optimizer (IGWO) with E-radius graph topology, GraphSAGE-derived node embeddings, and a Multi-Layer Perceptron (MLP) classifier. The E-radius graph enforces spatially constrained connectivity, while GraphSAGE facilitates the extraction of both structural and contextual node features. These features are subsequently input into the MLP for the classification and prediction of viable CH candidates. IGWO employs a multi-objective fitness function that simultaneously optimizes residual node energy, communication distance, network latency, Signal-to-Noise Ratio (SNR), and network connectivity metrics. The proposed methodology is validated on IIoT dataset, exhibiting enhanced energy efficiency, increased coverage ratio, and reduced latency in comparison to conventional clustering algorithms. Results substantiate that the integrated approach offers a scalable, adaptive, and energy-efficient solution for IIoT network orchestration.