Naslov
On behaviour of fuzzy norms and F learning automata in distributed multicriteria network routing

Autori
Lukač, Krešimir ; Lukač, Zrinka ; Tkalić, Mladen

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of the 21st IASTED International Multi-Conference on Applied Informatics / M. H. Hamza - Anaheim : Calgary : Zurich : ACTA Press, 2003, 686-691

Skup
IASTED International Conference on Parallel and Distributed Computing and Networks (PDCN´2002)

Mjesto i datum
Innsbruck, Austrija, 10.02.2003. - 13.02.2003

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
fuzzy logic; learning automata; multicriteria routing

Sažetak
In this paper we present a new model for distributed multicriteria dynamic routing and its behavior investigated through the series of experiments. This approach combines theory of learning automata with fuzzy logic theory. We call these automata the F type learning automata. Well known learning automata of P, Q and S types are special cases of this F type automata. We prove that these automata are strictly distance diminishing. Efficiency of the three combination of fuzzy t and s norms: min-max, product-algebraic sum and drastic product-drastic sum has been evaluated as well. Simulation results of the circuit switched telecommunication network obtained whereby two criteria: quality and price, have been taken into account simultaneously have shown superiority of product-algebraic combination of fuzzy norms under nominal and overloaded network conditions in comparison to another tested norms. The influence of a period in which F automaton updates its action probabilities based on fuzzy environment feedback on the network gain has been investigated as well.

Izvorni jezik
Engleski

Znanstvena područja
Matematika

POVEZANOST RADA