A Genetic Algorithm-Based Intrusion Detection System

In this paper, a novel approach to intruder detection is introduced. The approach, based on the genetic algorithms, improved detection rate of the host systems which has traditionally relied on known intruder patterns and host addresses. Rather than making judgments on whether the access is instrusion or not, the systems can continuously monitor systems with categorized security level. With the categorization, when the intruder attempts repeatedly to access the systems, the security level is incrementally escalated. In the simulation of a simple intrusion, it was shown that the current approach improves robustness of the security systems by enhancing detection and flexibility. The evolutionary approach to intruder detection enhances adaptability of the system.