Adaptive Trust-Based Intrusion Prevention Using Graph Neural Computing in MANET Environments

Abstract

Mobile Ad Hoc Networks (MANETs) have emerged as highly important communication infrastructures for enabling decentralized, self-organizing, and infrastructure-independent networking across dynamic wireless environments. MANET architectures are increasingly adopted in military communication systems, disaster recovery operations, emergency response coordination, intelligent transportation systems, IoT ecosystems, and autonomous distributed computing applications due to their flexible deployment and adaptive networking capabilities. However, the absence of centralized management, continuously changing topology, limited computational resources, and open wireless communication environments make MANETs highly vulnerable to security threats including black hole attacks, wormhole attacks, Sybil attacks, flooding attacks, selective forwarding, and malicious route manipulation. Traditional intrusion prevention mechanisms often rely on static trust evaluation, rule-based anomaly detection, and conventional routing security approaches that struggle to adapt to rapidly changing network conditions and evolving attack patterns. Recent advancements in graph-based machine learning and intelligent trust coordination provide promising opportunities for improving adaptive cybersecurity mechanisms in MANET environments. Graph Neural Networks (GNNs) enable contextual relationship learning among network nodes, communication paths, routing interactions, behavioral dependencies, and trust propagation mechanisms. Adaptive trust-based intrusion prevention mechanisms integrated with graph neural computing can continuously evaluate node behavior, dynamically update trust relationships, detect malicious communication patterns, and optimize secure route selection within highly dynamic network environments.