A Data Mining Approach for Predicting Reliable Path for Congestion Free Routing Using Self-motivated Neural Network

Congestion in computer networks is a significant problem due to the growth of networks and increased link speeds. Now it is common to see internet gateway drops 10 the incoming packets because of local buffer overflows. An optimal solution for this problem is Predicting congestion free path(s) by learning the dynamic characteristics of networks and its topology. The factors that influence the prediction of such path(s) have the characteristics viz., dynamic, non-linear, incertitude, etc., which make traditional data mining approach, like neural prediction have to process a large amount of convoluted data. In this paper, we proposed prediction model rather than mathematical model for finding congestion free path(s). We introduced a self-motivated learning in the training phase of an improved functional link feedforward neural network for predicting reliable path that offer congestion free path(s).