Link capacity control of ATM using genetic neural network algorithm

This paper explores the application of the genetic neural network (GNN) in the link capacity control of the asynchronous transfer mode (ATM). To solve the problem encountered in ATM network, use an artificial neural network (ANN) predictor to learn the relationship between offered traffic and performance of network from observed data automatically and predict the performance of network, and use GA to optimize the control. Computer simulation using two optimization object methods shows that this method can assign the link capacity correctly. Compared with traditional methods, the new way does not need complicated mathematical calculations and is easy to implement.