ChaTQ-DR: Chaos-Guided Trust-Aware Q-Learning with Dragonfly Optimization for Secure Routing in MANETs

Objectives: The decentralized structure of Mobile Ad Hoc Networks (MANET) makes them suitable for strategic communication, disaster recovery, and vehicular communication. yet no centralized monitoring system is in place, it makes the system extremely vulnerable to many types of routing attacks. Although the goal of current research is to enhance secure routing plans, the baseline and learning frequently use greedy exploration, which leads to false alarms in the detection rate. The searching exploration is carried out using the static trust rating model. This paper's primary objective is to avoid considering it as an independent model and instead concentrate on an adaptive and reliable routing method for trust calculation based on reinforcement learning. Method: A new Chaos-Guided Trust-Aware Q-Learning with Dragonfly Optimization (CTQ-DO) framework is put forth in this study. By enhancing the Dragonfly optimizer's exploration capabilities, chaotic perturbations help avoid premature convergence. The chaos-controlled Dragonfly update rule and a trust-modulated Q-learning equation. Q-value updates are directly infused with trust, allowing for reinforcement-based behavioural valuation, while explainability techniques make routing decisions transparent. Findings: Simulation results across varied mobility and attack scenarios demonstrate that CTQ-DO achieves up to 19% higher packet delivery ratio, significant reduction in routing overhead, and superior resistance to malicious nodes compared to traditional Q-Learning, AODV, DSR, and OLSR protocols. Novelty/Contribution: The proposed model uniquely combines chaotic metaheuristic exploration, trust-driven reinforcement learning, and explainability, establishing CTQ-DO as a secure, adaptive, and interpretable routing paradigm for MANETs. Keywords: Mobile Ad Hoc Networks (MANETs), Secure Routing, Chaos Theory, Trust-Aware Q-Learning, Dragonfly Optimization, Explainable Routing