CASPIAN JOURNAL
MANAGEMENT AND HIGH TECHNOLOGIES
MULTI-AGENT UNAPPROVED INFORMATION LEAK DETECTION SYSTEM IN E-LEARNING SYSTEMS
 |
Read | Prikhodko Maxim A. MULTI-AGENT UNAPPROVED INFORMATION LEAK DETECTION SYSTEM IN E-LEARNING SYSTEMS // Caspian journal : management and high technologies. — 2011. — №3. — pp. 113-118. |
Prikhodko Maxim A. - Cand. of Physics and Mathematics, Doctoral trainee, Moscow State Mountain University, 6, Leninsky avenue, Moscow, 119991, Russia, spex19@mail.ru.
This work considers problems of unapproved leaks in multi-agent informational communication systems. By unapproved leaks receiving of restricted data by regulated means is implied. The concept of unapproved leak is illustrated by the example of interactive knowledge control systems (Internet-testing), where lack of restriction may lead to exposure of the correct answers. Present methods of data restriction and their disadvantages are considered. Along with the problem of unapproved leak the problem of unapproved leak detection is formulated, that is particularly actual for the systems with huge number of users, requesting and receiving information in real-time (Internet-portals). The possibility of usage of multi-agent systems for solving the formulated problems is proved. As the criteria of the origin of unapproved leak it is proposed to use entropic characteristics of user’s actions and data flows of a system. Advisability of usage of multiagent systems, controlling data flows with restricted information on different levels of data abstraction is proved. Concept of multi-agent unapproved leaks detection and prevention system is listed. System requirements and current tasks are formulated.
Key words: leak,unapproved leak,e-learning system,multi-agent system,distributed system,agent,counter-agent,intellectual agents,rival agents.