Hybrid whale-gray wolf optimization for efficient intrusion detection in the Internet of Things

Abstract The extensive implementation of the Internet of Things (IoT) has transformed how mobile devices transmit data and facilitates communication and automation among interoperable items. Our lives and industries have recently evolved from smart household gadgets to industrial machinery with a heightened dependence on IoT devices. However, the recent rise in intrusions during data transfer via the Internet has also led to heightened security risks. Devices generate excessive information, allowing hostile individuals to exploit system vulnerabilities for illegal access, control of the device, and interruption of vital systems. Combatting these threats has become central to beefing up defenses for various systems. This research develops a novel hybrid optimization algorithm that blends the Whale Optimization Algorithm (WOA) and gray wolf optimizer (GWO) to strengthen intrusion detection in IoT environments. Attack detection is significantly enhanced by harnessing the global optimization capabilities of WOA in conjunction with the strong classification features of GWO. The WOA-GWO algorithm has been evaluated using multi-class and binary intrusion detection datasets, specifically Bot-IoT and NSL-KDD. Compared to prior techniques, the WOA-GWO model demonstrates superior classification accuracy (92.21% on Bot-IoT, 75.91% on NSL-KDD), precision (99.2% on Bot-IoT, 78.59% on NSL-KDD), F1 measure (99.29% on Bot-IoT, 77.47% on NSL-KDD), and recall (99.12% on Bot-IoT, 77.54% on NSL-KDD) highlighting its significant potential in enhancing IoT security.