Time series based forecasting methods in production systems: A systematic literature review

Abstract

Forecasting in production systems is used to anticipate their quality, efficiency, and yield. However, to the best of our knowledge, there exists no systematic review for industrial forecasting approaches. Thus, this work aimed to address this gap through a systematic literature review. The quantitative results revealed that industrial forecasting models are mainly applied in three economic sectors, with recurrent neural network models being the dominant approach. Moreover, this work proposes a classification of forecasting applications based on common characteristics found in reviewed sources. Several additional insights were produced, and future research directions were elaborated. Hence, this systematic review fosters an understanding of the current state-of-the-art of industrial forecasting approaches and facilitates future research initiatives