Hybrid Optimization Enabled Adaptive Quantum-Inspired Forward Taylor Net for Detecting Malicious Node in a 5G Mobile Network

In this work, the Chronological Addax Optimization Algorithm-based Adaptive Quantum-inspired Network (CrAOA-AQuNet) is proposed for malicious node detection. Initially Fifth Generation (5G) mobile network simulation is performed, and routing is carried out by Ad hoc On-Demand Distance Vector (AODV). Then, the trust is calculated for assessing the network reliability. Here, the framework integrates an Adaptive Hybrid Attention Network (AHANet), which focuses on critical network features and reduces noise to enhance true positive detection, and a Quantum-inspired Convolutional Neural Network (QuCNet), which captures complex malicious behavior patterns to improve robustness. Taylor series enhancements refine layer computations for precise feature representation, while the Chronological Addax Optimization Algorithm (CrAOA) optimizes network weights efficiently, ensuring faster convergence with improved performance. Experimental results demonstrate that CrAOA-AQuNet achieves superior performance, with an accuracy of 97.31%, a True Positive Rate (TPR) of 96.96%, and a True Negative Rate (TNR) of 96.60%, outperforming existing methods. Here, the performance improvement gained by the proposed CrAOA-AQuNet concerning accuracy metrics is 9.826%, 8.843%, 7.628%, 5.997%, 5.057%, and 2.337% higher compared to the existing models. These results demonstrate the framework’s robustness, scalability, and effectiveness in securing 5G networks, providing a reliable solution for real-time malicious node detection.