A Multi-Agent Deep Reinforcement Learning Framework for Detecting and Mitigating DDoS Attacks in IoT Networks
Bhushan Bharat Shaharkar, Tanusha Mittal, Krishna Chaitanya Sunkara, Vishal Shukla, Ishan Desai, Ujal Kumar Mookherjee · 2024
The rise of Internet of Things (IoT) networks has introduced new opportunities for innovation but also increased vulnerability to cyberattacks, particularly Distributed Denial-of-Service (DDoS) attacks. Traditional detection and mitigation mechanisms often struggle to handle the dynamic and distributed nature of IoT systems. In this study, we propose a Multi-Agent Deep Reinforcement Learning (MADRL) framework designed to detect and mitigate DDoS attacks in IoT networks in real-time. The framework consists of multiple collaborative agents, each deployed at different network layers or nodes, trained using deep reinforcement learning to identify attack patterns and take immediate mitigation actions. Through cooperative learning, these agents adapt to evolving attack strategies and optimize defense policies. Experimental evaluations demonstrate that the proposed MADRL framework outperforms traditional techniques by achieving higher detection accuracy, faster response times, and reduced false positive rates. Our results indicate that this framework can effectively secure IoT networks against complex DDoS attacks, maintaining network stability and availability.