Evolutionary gravitational neocognitron neural network espoused blockchain-based intrusion detection framework for enhancing cybersecurity in a cloud computing environment

Abstract

Cloud computing offers scalable, on-demand resources but remains highly vulnerable to cyberattacks, where the nonlinear and dynamic nature of network traffic makes detection especially challenging. This study introduces EGNNN-GROA-BGPoW-IDS-CC, a novel intrusion detection framework that explicitly addresses the nonlinear complexities of attack patterns by integrating deep neural modeling, evolutionary optimization, and blockchain technology. Central to the system is the Evolutionary Gravitational Neocognitron Neural Network (EGNNN), capable of learning nonlinear feature hierarchies, and optimized using the GarraRufa Fish Optimization Algorithm (GROA) for enhanced detection accuracy. Input data from the NSL-KDD dataset are preprocessed using the Developed Random Forest with Local Least Squares (DRFLLS) to reduce noise and nonlinear redundancy, followed by feature selection through the Dynamic Recursive Feature Selection Algorithm (DRFSA) to capture the most influential nonlinear dependencies. For secure alert logging, a Blockchain-based Green Proof of Work (BGPoW) ensures lightweight, tamper-proof consensus while maintaining energy efficiency. Implemented in Python, the proposed model demonstrates superior performance, outperforming state-of-the-art systems such as BiLSTM-DBF-IDS-CC and DBN-ResNet-IDS-CC, with accuracy improvements of 32.76% and 15.78%, respectively. Overall, EGNNN-GROA-BGPoW-IDS-CC presents a high-performance, energy-efficient solution that explicitly addresses the nonlinear behavior of cyber threats, thereby advancing sustainable cybersecurity in cloud environments.