Mestrado em Engenharia de Produção

URI Permanente para esta coleção

https://repositorio.ufcat.edu.br/handle/tede/84

Navegar

Navegando Mestrado em Engenharia de Produção por Assunto "Aprendizado de máquina"

Agora exibindo 1 - 1 de 1
  • Imagem de Miniatura
    Item
    Previsão do lead time de processos usando mineração de dados
    (Universidade Federal de Goiás, 2021-03-09) Oliveira, Maíza Biazon de; Cordeiro, Douglas Farias; http://lattes.cnpq.br/5269312530540199; Silva, Núbia Rosa da; http://lattes.cnpq.br/8448585265537772; Silva, Núbia Rosa da; Duarte, Carlos Antonio Ribeiro; Silva, Sergio Francisco da; Cordeiro, Douglas Farias
    The era of Industry 4.0 leads to constant adaptations of production processes and generates a significant amount of information. In this way, information management becomes a crucial factor to guarantee the competitive strategy in the industries. One of the information to be managed is textit lead time, time between the customer requesting an order and it being available. Usually, it can be estimated using expensive measurements or traditional methods that do not normally reflect the actual behavior of the data or do not support the significant amount of information generated in Industry 4.0. In addition, there are gaps in the literature on textit lead time forecasting, such as the use of smart methods to predict textit lead time across the supply chain. In this context, the objective of this research is to use data mining using machine learning algorithms to predict the textit lead time in real processes. The proposed methodology made use of the textit Knowledge Discovery in Databases (KDD) cycle structured in the selection, pre-processing, transformation, data mining and knowledge discovery phases. The learning algorithms for textit Linear Regression (LR), textit Random Forest (RF), textit Support Vector Machine (SVM), textit K-Nearest Neighbors (KNN) were tested and textit Multilayers Perceptron (MLP). To validate the experiments, three databases from the Electronic Information System (SEI), a supply chain from a pharmaceutical logistics sector and from the industrial automation sector for the ceramic sector, were used. The results showed that data mining is an effective tool for analyzing data generated in the fourth industrial revolution for forecasting textit Lead time and decision making on production planning and control.