Abstract
It is estimated that Fixed Telephony Service will have an important share in a near future, just like in the past. This paper discusses the “artificial neural networks”, which can correctly model non-linear multi-input dynamic systems, and forecasts the number of the subscribers of this service, comprising a period of future 5 years by using a solution searching technique used in simulation applications called “genetic algorithms”.
Keywords:
Artificial neural networks, genetic algorithms, number of fixed telephony subscriber, forecast model


