期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2015
卷号：3
期号：6
DOI：10.15680/ijircce.2015.0306142
出版社：S&S Publications
摘要：Continuous user authentication is an important prevention-based approach to protect high securitymobile adhoc networks (MANETs). Intrusion detection systems (IDSs) are also significant in MANETs to identifymalicious activities. Considering these two approaches mutually is an effective in optimal security design to protecthigh security MANETs. User authentication and authorization desires to be performed continuously and frequently,since the chance of a device in a hostile environment being captured is extremely high. To obtain the optimal scheme ofcombining continuous authentication and IDSs in a distributed manner, we formulate the scheduling problem as aPartially Observable Markov Decision Process (POMDP) multi-armed bandit problem. We present optimal Gittinsindex policy to solve the scheduling problem for a huge network. The Gittins index of a process is a function of thatprocess’s characteristics and its information state. Value iteration algorithm is used for computing Gittins index. Foreach process, information state and Gittins rewards are calculated. Based on the information states and the rewards, theprocesses are scheduled using POMDP policies. The optimal policy is to select the process with the largest rewardGittins index. The process with the higher probability of being in the better state has a higher possibility of beingchosen at that time slot. The optimal policy picks the authentication process or the intrusion detection system with thelargest Gittins rewards.
关键词：Biometric traits; Continuous authentication; intrusion detection system; mobile adhoc network;POMDP.
