Abstract
Attack detection is one of the most important issues for computer networks security. Intrusion Detection System (IDS) monitors network for malicious activity. In this paper, a method is proposed to identify the attacks in IDS. This method generates rule sets for Intrusion Detection System using non-dominated sorting genetic algorithm (NSGA-II). NSGA-II is one type of multi objective genetic algorithms. This method considers features of computer connection and defines two different fitness functions for generating the rules. The advantage of this method compared with previous methods which applied Evolutionary Algorithm. Since some methods applied one fitness function or convert many objectives to single objective, they lost many features.