AI-Driven Dynamic Firewall Optimization Using Reinforcement Learning for Anomaly Detection and Prevention

Abstract

The growing complexity of cyber threats has rendered static firewalls increasingly ineffective for dynamic, real-time intrusion prevention. This paper proposes a novel AI-driven dynamic firewall optimization framework that leverages deep reinforcement learning (DRL) to autonomously adapt and update firewall rules in response to evolving network threats. Our system employs a Markov Decision Process (MDP) formulation, where the RL agent observes network states, detects anomalies using a hybrid LSTM-CNN model, and dynamically modifies firewall configurations to mitigate risks. We train and evaluate our framework on the NSL-KDD and CIC-IDS2017 datasets using a simulated software-defined network environment. Results demonstrate significant improvements in detection accuracy, false positive reduction, and rule update latency when compared to traditional signature- and behavior-based firewalls. The proposed method provides a scalable, autonomous solution for enhancing network resilience against complex attack vectors in both enterprise and critical infrastructure settings.