Abstract: Mobile agent environment is a composition of migrating programs between hosts. These programs (mobile agents) provide services for hosts but they need to fairly distribute themselves throughout host machines. A market strategy could be used to regulate the action of mobile agents among the network. Prices are virtually set to derive the allocation of resources to agents. This of course affects demands of agents to resources and so their utility driven from them. This paper proposes a new strategy in making a fair mobile agent distribution among the network hosts. This is done through the adaptive function of utility driven distribution for mobile agent scheduling. A prototype model is synthesized to simulate the network attitude. This was done using an equation that dynamically adjusts to the current state of the distributed agents among the network servers (hosts). A discussion about the performance of the network is provided.
Abstract: Security in mobile ad hoc networks (MANETs) is difficult to achieve, notably because ad hoc networks are vulnerable to attacks due to the dynamically changing topology, distributed nature, lack of infrastructure and the lack of a centralized monitoring or management point. In this paper, we present a novel intrusion detection system (IDS) by using mobile agents over cluster-based routing protocol (CBRP) for MANETS. The reliance on mobile agents for IDS is desirable due to lightweight computation, reduction in network bandwidth usage by moving data analysis computation to the location of the intrusion data, support of heterogeneous platforms, and flexibility in creating distributed IDS for clustered mobile ad-hoc network. The proposed IDS is evaluated through the use of the ns-2 network simulator. Preliminary performance evaluation results illustrate the efficiency of the proposed IDS in terms of attack detection rate and percentage of false alarms.
Abstract: Reaching a testing system that satisfies a student ability level has been a study issue for a long time. Computerized adaptive testing (CAT) is one of the solutions for measuring student ability giving him a test that suit his ability. But CAT has some limitations that prevent it from the perfections that the students are aiming. This research will solve some of these limitations hoping to reach a better system performance of CAT and a more accurate student ability measure. The limitation is CAT systems always testing the student in random questions pattern in the whole topic so, student not administered enough items of all concepts in the topic failing to insure that their final ability estimate is accurate enough. This paper solved that the student couldn’t really understand a concept in a topic end by bad results that he may get. Offering a solution of this limitation using software agent technology with CAT.