期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2016
卷号：88
期号：1
出版社：Journal of Theoretical and Applied
摘要：Mobile ad hoc network (MANET) is a dynamic and promising research domain that contributes to the development of wireless networks. MANET characteristics such as dynamic topology, open media access, wireless communication, and resource limitations introduce several security threats. The MANET security is then Vulnerable. The use of traditional security techniques cannot be directly applied for the case of MANET, hence the need to develop more suitable methods and security algorithms. Artificial immune systems have been widely used in the field of MANET security. Several immune algorithms were used and implemented as intrusion detection systems (IDS). This paper gives an overview of major immune algorithms like Negative Selection Algorithm, Clonal Selection, and the danger theory. It then describes our approach to implement an IDS based on the combination of these immune algorithms to better simulate the human defense mechanism. The proposed algorithm is named �Combined Immune Theories Algorithm� (CITA), and is implemented with the AODV routing protocol. CITA is designed to be an auto evolutionary algorithm with some learning capabilities in order to be able to recognize and exclude unknown intruders based on deviations from normal behaviors. Simulations are done using Network simulator 2 (NS2) with increasingly insertion of malicious nodes running Resource consumption attacks. Simulation results show improved performance of the proposed CITA algorithm compared to SAODV routing protocol in terms of better detection rates, low false positive detection rates, higher throughput and packet delivery ratio.
