Effect of change in rate of genetic algorithm operator on composition of signatures for misuse intrusion detection system

Intrusion detection (ID) has become an important area of research since building a system with no vulnerabilities has not been technically feasible. In this paper, effect of change in rate of genetic algorithm operator on fitness value in composition of signatures for misuse intrusion detection system is presented. The proposed genetic algorithm uses a set of classification rules which are generated from a predefined intrusion behavior. From the results it could be concluded that if crossover performed is not sufficient, there is not sufficient sharing of genes. If we crossover too much, good segments of individuals get split up a lot. This allows some individuals with high fitness's to be copied directly to the next population.