Uso de conhecimento teórico e de especialista para previsão de demanda.

• Autor(a) principal: Dias, Andreza Silva
• Data de Publicação: 2004
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Repositório Institucional da UFSCAR
• Texto Completo: https://repositorio.ufscar.br/handle/ufscar/3546
• topic: Planejamento da produção; Previsão de demanda; Inteligência artificial; Sistemas especialistas (computação); Ajuste de previsão; Forecasting; Expert systems; Forecasting adjustment; Artificial intelligence; ENGENHARIAS::ENGENHARIA DE PRODUCAO

In a highly competitive market as the actual one, the production management must ensure that goods and services are supplied to the customers in right time, so they can offer a competitive advantage or, at least, can be sold by a price that surpasses the production costs. It can be affirmed that one of the variables which determines the service level, to the customer, is the product availability, such in rapidity terms, as the amount of the available product. A determinant function for the availability of a product or service is the forecasting, it analyses the records about the past sales and it infer the possible behavior of the consumers for the following periods. The demand is influenced by qualitative and quantitative factors that are not considered simultaneously by the traditional forecasting methods. Thus, observing the trends in the use of expert systems and the importance of the demand forecast, this research aims to develop an expert system for demand forecast that combines qualitative and quantitative methods. To test and to analyze the performance of the systems a case study in an industry of soft drink had been carried through. This work considers an adaptation to the method of quantitative forecast adjustment with judgment to develop an expert system for demand forecast. The quantitative forecast is outcome by the combination of the forecasts generated for the most known methods. The results obtained in the tests carried through with data that had been collected during the study case confirm that the use of judgment as adjustment form of forecasting obtained by quantitative methods improve the quality of the forecast. Moreover, this work allowed proving that the knowledge of an expert who calculates the forecasting using his own judgment allows the construction of a Knowledge Base of an Expert System.