SWARM INTELLIGENCE FOR REAL-TIME INTRUSION DETECTION:A BIO-INSPIRED FRAMEWORK USING ANT AND BEE COLONY OPTIMIZATION"

Abstract

As these threats become more advanced and new ones appear, relying on constant threat detection is more important for network security. Traditional IDS often have problems adapting and reacting quickly to large and diverse environments. This document discusses a bio-inspired design for IDS that uses Ant Colony Optimization (ACO) and Bee Colony Optimization (BCO) algorithms to perform real-time threat scanning. Focusing on how natural swarms are decentralized and self-organized, the system models network traffic analysis as a combined effort to detect anomalies while locating sources. A new framework is built that eases the routing of agents, increases the importance of selected features and improves detection accuracy while using few resources. Tests done using the NSL-KDD and CICIDS2017 datasets have shown that the bio-inspired IDS achieves high accuracy, few false alarms and better ability to adapt than classical machine learning models. We have found that using swarm intelligence is a suitable and scalable way for building better IDS, fit for protecting systems in modern, active cybersecurity settings.