Generating Various Patterns of Intrusion using IGA

As the computer environment changes significantly, several mechanisms such as firewall are developed and intrusion detection system (IDS) is the representative among them. However, because many IDSs are modeled and evaluated with known intrusion patterns, they have inertia not to detect the intrusions which are unknown or transformed. In this paper, we propose an IGA-based method to generate the novel intrusion patterns. The fitness values of generated intrusion patterns are evaluated by users, and we apply the generated intrusion patterns to LinSTAT, one of the most famous misuse detection systems. As a experimental results, various and available intrusion patterns are generated about 73%.