Misuse Detection via a Novel Hybrid System

Intrusion detection systems (IDS) are tools located inside computer networks that analyze the network traffics. In this paper, a novel fuzzy-evolutionary system is presented to effectively detect the intrusion in computer networks. This system utilizes a hybridization of simulated annealing heuristic and tabu search algorithm to improve the accuracy of fuzzy if-then rules as intrusion detectors. Each of these algorithms has its advantageous and disadvantageous. Using the hybrid model of both algorithms, the proposed system employs the good features of them to improve the accuracy of obtained rules. Evaluation of the proposed system is done on the KDDCup99 Dataset which has information about normal and intrusive behaviors in networks. Results of our model have been compared with several well-known intrusion detection systems.