Especificação de uma ferramenta de previsão de demanda para equipamentos de infra-estrutura de telecomunicações
AUTOR(ES)
Sacha Tadeu Branco
DATA DE PUBLICAÇÃO
2008
RESUMO
Over the past decades the telecommunication market hás been more and more deregulated and competitive, while more and more telecommunicatios techonologies are available for the users. Companies operating in these markets have to rely on demand forecasts to justify their level of investmants and ensure the capacity availability at the right time. With so many variables in this market, forecasting is becoming more challenging. This work presents a quantitative demand forecast model suitable to the market of telecommunication infrastructure equipments. It begins with the research of demand forecasting and shows in which step of a forecasting process forecast tools shall be used. Through na analysis on the characteristics of some of time series, causal and artifical intelligence forecasting tools, a neural network forecast model is recommended to be the appropriate method for accurate forecast. After a research and the definition of the inputs, ouputs and parameters, forecast simulations are performed and compared to the forecast quality method currently obtained by the company under study. The results show that the use of neural networks approach hás a better performance than the current company forecast method and also that it hás a high potential usage for this kind of market.
ASSUNTO(S)
engenharia de producao neural networks (computer science) telecommunication telecomunicações - legislação production engineering telecomunicações - equipamento e acessórios telecomunication engenharia de produção - dissertações redes neurais (computação)
