CONGESTION CONTROL MODELLING BASED ON LOCALLY LINEAR NEUROFUZZY NETWORK

In this paper, a sample of a congestion control data will model by neuro-fuzzy algorithm. To build the neuro-fuzzy model, a locally linear learning algorithm, namely, Locally Linear Mode Tree (LoLiMoT) is used. Then, a congestion controller is applied to the identified model This intelligent algorithm provides more speed, less training time and less sum square error in simulation than MLP. Simulation will done with some cell loss data that are fetched from a broadband integrated services digital network(B-ISDN), and represents not only maximizing in speed but also make less sum square error in optimization of parameters.

کلیدواژه ها:

Congestion Control ، Computer networks ، Neuro-fuzzy networks ، Locally linear models ، LoLiMoT