Membership Details (Research Institute)
Contact: Gorodetsky, Vladimir
Member node: 082
39, 14-th Liniya, 199178, St.Petersburg, Russia
Russian Federation
>>> Include yourself at this organisation
>>> Update your name or personal details listed below
>>> Submit papers, projects etc to the archive
>>> Update Member Description or Contact details
Description
The Institute researchers work in different aspects of multi-agent system and multi-agent technology since 1995. The main research topics covering multi-agent scope are as follows:
(1) multi-agent system methodology and design, implementation and deployment technology; (2) software tool supporting the above methodology and technology; (3) agent-based modeling and simulation and (4) design, implementation and deployment of applied multi-agent systems in various problem domains.
In particular, during last six years three versions of Multi-agent System Development Kit (MASDK) were developed and used for design, implementation and deployment of applied multi-agent systems prototypes. Among such applications, the following software prototypes were developed and implemented:
-Multi-agent intrusion detection system deployed over four hosts of LAN and comprising 28 agents of different destinations collaborating through KQML communication language while exchanging messages in XML.
-Multi-agent system for distributed data mining and decision making (information fusion) that was used for development of several software prototypes, for example, based on KDDCup-99 case study and also based on case study developed in the Institute. The last system is destined for distributed learning and decision making in intrusion detection scope. It uses use 12 different sources of information of a host, in which the traces of intrusions are manifested.
-Multi-agent planning and scheduling system for transportation logistics based on well known Solomon%2's benchmark. This software prototype is considered as a starting point for design and implementation of multi-agent systems for the problem "mission planning and logistics". This software system was a finalist of agent-oriented software competition organized within 3d Agent Information Day (Barcelona, 3-10 February, 2003).
-Agent-based modeling and simulation of distributed attacks against computer network. This software prototype simulates activity of hackers that are capable to implement about 120 different attacks, represented as attack ontology. This software tool was highly evaluated within computer network scientific community (see item 2 of the selected publications given below) because it is considered as a software prototype for automatic rapid validation of computer network security policy.
The latest theoretical research concerns with formal specification languages and formal verification (model checking) in the area of multi-agent intrusion detection systems and in the future it is intended to generalize this research to verification of protocols in open multi-agent systems.
The Institute was an initiator of the currently well known series of conferences of Central and Eastern Europe on Multi-agent system (CEEMAS-99, St. Petersburg, Russia), which is successfully continuing now (CEEMAS-01 in Krakow, Poland; CEEMAS-03 in Prague, Czech Rep.; CEEMAS-05 to be held in Budapest, Hungary).
Now Institute works with promotion of new series of International Workshops "Autonomous Intelligent Systems: Agents and Data Mining" devoted specifically to proactive computing. The first workshop of this series is scheduled on June 5-9, 2005, Call for papers will be disseminated in May, 2004.
Several Ph.D. students in the Institute are working with the problems from multi-agent scope.
The Institute researcher have rich experience in international collaboration on multi-agent systems. Among its Partners are such institutions and companies, as Fraunhofer First (Germany), Cadence (USA), US Air Force Research Laboratory (IF directory) and others. Institute is a participant of STREP Project "POSITIF" within Six Framework Program of European Commission (2004-2006)in which, although the Project is devoted to computer network security problems, the researchers of the Institute are responsible for agent-based validation of computer network security policy.
Insitute was a member of both AgentLinkI and AgentLink II.
This list was generated on Mon Sep 10 16:39:02 BST 2007. |
