Optimizing Network Energy In Cloud-Assisted IOT With Wireless Sensor Networks Using Similarity-Navigated Graph Neural Networks

Wireless Sensor Networks (WSNs) have been identified as important in cloud assisted internet of things (IoT) applications such as remote monitoring and energy saving. However, energy consumption and having limited network life time are still the major challenges in WSNs. Conventional methods of energy optimization does not achieve an optimal balance between avoiding overhead and effective communication. To overcome these drawbacks, this work presents a new Similarity-Navigated Graph Neural Network with Running City Game Optimizer (SNGNN-RCGO) model for energy optimization in IoT-WSN structures. It has a wireless sensor layer at the application tier, interconnecting layer at the middle tier, and the cloud layer at the software tier. SNGNN also reduces the energy usage by grouping nodes with similar nodes but excludes low energy node in the group. RCGO enable SNGNN hyperparameters tuning by exploiting exploration-exploitation trade-off while GBO enables efficient route and cluster head selection. In this regard, the proposed SNGNN-RCGO finds the highest residual energy (361.32 units) for node and network life time up to 1809 rounds with 195 alive nodes proving better energy efficiency. It also reduces the overall network overhead to 2 units and realizes the smallest end-to-end delay of 41.3 ms, which explains its efficiency and effectiveness for IoT and WSN. The proposed approach thus eliminates shortcomings of prior techniques since it enhances network energy efficiency and network lifetime with little added cost and complexity.