Detecting Zero-Day Attacks Using Deep Learning with Pelican Optimization Algorithm in IIoT Environments

5G arises as the base for the Industrial Internet of Things (IIoT); it enables the unified, low-latency hybrid of cloud computing and Artificial intelligence (AI), thus strengthening the complete industrial process within a structure of intelligent and smart IIoT environments. Simultaneously, the constantly evolving landscape of cybersecurity hazards in the Internet of Things (IoT) domain presents opportunities for enhanced safety complexities. Recognizing zero-day threats is a challenging task due to the indefinite nature of security exposures. This study proposes a new Metaheuristic Optimization Algorithm with Deep Learning Enabled Zero-Day Attack Detection (MHOA-DLZDAD) method for IIoT frameworks. The MHOA-DLZDAD method automates and effectively detects zero-day attacks. Initially, the MHOA-DLZDAD model undergoes min-max scalarization using data pre-processing to convert actual data into a suitable format. Moreover, the Elman Recurrent Neural Network (ERNN) method is utilized to detect zero-day attacks. Furthermore, the Pelican Optimization Algorithm (POA) method is employed for tuning the parameters. The experimental analysis of the MHOA-DLZDAD approach is conducted on a benchmark dataset, and the comparison study reveals a higher accuracy of 99.56% compared to other studies.