Improvement of grey prediction models and their usage for energy demand forecasting

Ervural B. C., Ervural B.

JOURNAL OF INTELLIGENT & FUZZY SYSTEMS, vol.34, no.4, pp.2679-2688, 2018 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 34 Issue: 4
  • Publication Date: 2018
  • Doi Number: 10.3233/jifs-17794
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.2679-2688
  • Istanbul Technical University Affiliated: Yes


Optimal energy planning is one of the most significant issues for all over the world, in short, medium and long term strategic projections of countries due to the vagueness and concerns about energy reliability and sustainability in limited resources. The dynamic and chaotic nature of the energy systems requires a well-constructed and multidimensional prediction model to create an urgent energy requirement planning. In this study, grey prediction models based on genetic algorithm (GA) and particle swarm optimization (PSO) are proposed to provide more realistic and quick energy demand forecasting with high accuracy. The grey modelling is a popular approach that can be used to construct a model with the limited sample of historical data. GA and PSO are used for the tuning of an optimal set of structural parameters of classical grey prediction model to obtain more robust and efficient solutions with minimum prediction errors. A case study using the data of Turkey is presented. Results confirm that the proposed methods demonstrate superior forecasting performance, compared with alternative models.