Intelligent agents for testing the internet

• Autorzy: Dziubich T. , Krawczyk H.
• Konferencja: 19th IMEKO TC-10 International Conference on Technical Diagnostics. Integration in Technical Diagnostics (22-24.09.1999 ; Wrocław)
• Języki publikacji: EN

Abstrakty

EN

The paper describes some kind of intelligent agents able to observe behaviour of both nodes and communication links in a computer network. Because network is dynamic, the behaviour of both nodes and variables changes frequently. The goal is to recognize changes which are important in the context of network behavioral anomalies that correspond to network faults. The processed measurement information is combined into high level measures which are intelligent agent's outputs. Based on the measures typical faults can be recognized and singnalized to fault management centre. The paper shows how to learn agents by the Bayesian rule and how to organize diagnosis on the system level.

Słowa kluczowe

EN

Bayesian networks; network diagnosis; intelligent agents

• Wydawca: Oficyna Wydawnicza Politechniki Wrocławskiej
• Czasopismo: Prace Naukowe Instytutu Górnictwa Politechniki Wrocławskiej. Konferencje
• Rocznik: 1999
• Tom: Vol. 86, nr 24
• Strony: 118--123
• Opis fizyczny: Bibliogr. 7 poz., rys. 3

Twórcy

autor

Dziubich T.

autor

Krawczyk H.

• Technical University of Gdańsk, 11/12 Narutowicza st., 80-952 Gdańsk, Faculty of Electronic, Telecommunication and Computer Science