Augmented Cyber-Physical Systems for Supply Chain 4.0 DDoS Attack Detection Deploying a TSO-DBN Deep Learning Approach Evaluation

Implementing robust detection methods is essential in the evolving realm of Supply Chain Risk Management 4.0, as cyber-physical systems (CPS) face increasing vulnerability to Distributed Denial of Service (DDoS) attacks. This paper introduces a cutting-edge approach to detecting DDoS attacks by integrating various artificial intelligence (AI) and data processing techniques. The process begins with Z-score normalization for pre-processing, ensuring consistent feature scaling. Next, Principal Component Analysis (PCA) is used for dimensionality reduction while retaining significant variation. To enhance detection performance, we optimize the feature subset using Tunicate Swarm Optimization (TSO) for effective feature selection. These selected features are then utilized by a Deep Belief Network (DBN), a deep learning model, to classify anomalies. The DBN leverages learned patterns to distinguish between normal and attack scenarios. To assess the effectiveness of the detection system, we evaluate performance indicators such as accuracy, precision, recall, and F1-score, ensuring the model's reliability. Additionally, the research examines the impact of various feature selection algorithms on the overall performance of the DDoS detection system to identify the most effective approach. This comprehensive strategy significantly enhances the resilience of cyber-physical systems against attacks, achieving a 99% success rate and improving the accuracy and reliability of DDoS detection in the dynamic landscape of Supply Chain Risk Management 4.0.