Use of artificial neural networks as support for energy saving procedures in telecommunications

In recent times, there has been great interest in the field of energy saving applied to a wide range of industrial reality such as production plants, services, public transportation companies and telecommunication companies. The energy consumption in a telecommutation central could raise interesting values, for this reason there is a great interest in applying forecasting techniques at this problem, in fact one of the major issues is related to the power transportation from the electrical production plant to the telecommutation devices, this task require an estimation of the next hours consumption that will be used as a base for the commercial agreement with the production partner, misestimating the future consumption could generate great losses of money. The authors applied successfully neural networks using both fully connected feedforward backpropagation (BP) networks and also Boltzmann machine's (BM) nets in order to obtain more reliable forecast. Particularly the BM offers a range of validity estimation that is very useful for what if analysis and simulation based decision support systems (DSS).