A hybrid GA-DL framework for energy-efficient clustering and lifespan extension in RWSNs

The limited energy capacity of sensor nodes remains a central constraint on the operational longevity of Wireless Sensor Networks (WSNs)-a challenge that is more critical in Rechargeable Wireless Sensor Networks (RWSNs), where sustaining long-term functionality depends on delaying the First Node Death (FND), a key point that often signals the beginning of network problems. While prior research has extensively explored heuristic algorithms coupled with energy-aware routing mechanisms to reduce transmission energy consumption and extend network lifetime, these methods typically require repeated iterations to converge. Such computational demands render them unsuitable for real-time applications. To overcome these limitations, we introduce a hybrid optimization framework that integrates Genetic Algorithms (GA) with Deep Learning (DL). This framework can run quickly to improve energy efficiency and extend network lifetime in RWSNs. The proposed approach builds upon an energy-aware clustering architecture that incorporates a lightweight version of the Residual Neural Network (ResNet) model. This clustering strategy promotes a more balanced distribution of energy consumption across the network, delaying early node failures caused by localized energy depletion. Based on this setup, we create an optimization scheme using a novel version of GA that helps the ResNet model-simplified for use on regular CPUs-learns an optimal CH selection policy. The GA component employs a custom fitness function designed to maximize the time to FND latency. The synergy between the clustering method and the GA-DL hybrid selection mechanism enables rapid and effective CH assignment without the iterative overhead typical of heuristic algorithms. Our experimental results demonstrate that the proposed scheme not only reduces energy consumption and extends the time to FND but also achieves near real-time performance. This makes it a promising alternative to traditional heuristic-based approaches, particularly in latency-sensitive RWSN applications. Together, these contributions mark a significant step toward the development of scalable, efficient, and resilient RWSN architectures.