Adaptive Whale Optimized Neural Intelligence for Cyber Attack Detection in Industrial IOT Environments

Abstract

Important cyber-physical systems now have a much larger attack surface due to the rapid
growth of Industrial Internet of Things (IIoT) infrastructures, making them vulnerable to sophisticated
and constantly changing cyberthreats. High-dimensional traffic data, class imbalance, and changing
attack patterns that are typical of IIoT contexts are common problems for traditional intrusion detection
systems (IDS). To address these issues, this research presents an Adaptive Whale-Optimized Neural
Intelligence (AWONI) model designed to identify cyberattacks in IIoT networks with accuracy and
efficiency The Whale Optimization Algorithm (WOA) and an Artificial Neural Network (ANN) are used
in the proposed technique to enable intelligent feature selection and optimize classification. While the
ANN classifier successfully models complicated non-linear connections for multi-class attack detection,
WOA is used to discover the most important characteristics within large-scale IIoT traffic data, reducing
dimensionality and boosting leaming efficiency. The CIC-IIoT-2025 dataset, which includes a range of
real-world industrial attack situations and typical traffic patterns, is used to evaluate the architecture.
According to experimental data, the proposed model outperforms conventional machine learning (ML)
and Deep Learning (DL) techniques and non-optimized neural approaches in terms of detection
accuracy, false alarm rates, and computing efficiency. To promote the development of robust and
intelligent industrial cybersecurity solutions, the suggested framework's adaptive, nature-inspired
architecture makes it ideal for implementation in real-time and resource-constrained IIoT security
situations.