Improved Agent Model for Network Security Evaluation Based on AIS

Abstract

This paper proposes a novel distributed agent model for network security evaluation. The method, which uses antibody concentration to quantitatively describe the degree of intrusion danger, is demonstrated. A new network security evaluation method using antibody concentration to quantitatively analyze the degree of intrusion danger level is presented. The method which uses antibody concentration to quantitatively describe the degree of intrusion danger is presented. Additionally, the hierarchical and distributed management framework of the proposed model is built, avoiding neglecting uncertain factors that the traditional method often does. Our experimental results show the model which enhances detection efficiency and assures steady performance in the ability of intrusion detection.