Communication network planning using artificial neural networks

A strategic plan of a digital communication system for one of the most underdeveloped areas of the world is described in the paper. Communication network planning using the usual computer programs was not effective. A trainable system of artificial neural networks is shown to have advantages in this unusual situation. A neural network switch used for the dimensioning of trunks is shown to give an economical solution to this problem when compared with the estimate by various well known vendors.<>