Network Intrusion Detection Method Based on Improved Simulated Annealing Neural Network

Abstract

Aimed at the intrusion behaviors are characterized with uncertainty, complexity, diversity and dynamic tendency and the advantages of neural network, an intrusion detection method based on improved simulated annealing neural network (ISANN) is presented in this paper. First the simulated annealing algorithm with the best reserve mechanism is introduced and it is organic combined with Powell algorithm to form improved simulated annealing mixed optimize algorithm, instead of gradient falling algorithm of BP network to train network weight. It can get higher accuracy and faster convergence speed. We construct the network structure, and give the algorithm flow. We discussed and analyzed the impact factor of intrusion behaviors. With the ability of strong self-learning and faster convergence of ISANN, the network intrusion detection method can detect various intrusion behaviors rapidly and effectively by learning the typical intrusion characteristic information. The experimental result shows that this intrusion detection method is feasible and effective.