Short-Term Load Forecasting For Energy Markets

  • Witold Bartkiewicz
  • Bożena Matusiak
Conference paper
Part of the Advances in Soft Computing book series (AINSC, volume 19)


The problem of the application of neural predictors for Short-Term Load Forecasting (STLF) for energy transactions planning in utility is discussed. The conditional expected value of the energy demand approximated by the neural network forecast does not necessarily indicate the optimal size of the order, according to the financial regulations associated with energy markets. Two approaches to the solution of this problem are presented. The first one is based on GANN — hybrid neural-genetic STLF system, trained with the financial cost function. The second approach rely on determining of the sources of uncertainty for classical neural predictor.


Energy Demand Energy Market Load Forecast Distribution Company Neural Network Output 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Bardzki W., Bartkiewicz W., Gontar Z., Zielinski J.S. (1998) Short-term electrical load forecasting in transformation period (Case Study). Proc. of the Intl. ICSC/IFAC Symp. on Neural Computation (NC ‘88), Vienna, Austria, Sept. 22–26, 324–328Google Scholar
  2. 2.
    Bartkiewicz W. (2000) Confidence intervals prediction for the short-term electrical load neural forecasting models, Elektrotechnik und Informationstechnik. 1 (117), 8–12Google Scholar
  3. 3.
    Bartkiewicz W, Gontar Z., Zielinski J.S. (2000) Neural-Heuristic approach to short-terni load forecasting problems. Proc. of the II International ICSC Symp. on Neural Computation, NC’2000, Berlin, Germany, May 21–26, 740–744Google Scholar
  4. 4.
    Bartkiewicz W. (2000) Distribution of the residuals of neural predictor for short-term electrical load forecasting problems, Proc. of the V Polish Neural Networks Society Conf.: Neural Networks and Soft Computing, Zakopane, June 6–10, 112–117Google Scholar
  5. 5.
    Bartkiewicz W., Gontar Z., Zielinski J.S., Bardzki W. (2000) Uncertainty of the short-term load forecasting in utilities. Proc. of the Intl Joint Conference on Neural Networks (IJCNN 2000), Como, Italy, July 24–27, vol. VI, 235–240Google Scholar
  6. 6.
    Bartkiewicz W. (2001) Error bars for short-term load forecasting neural networks models. Proc of X-th Intl. Conf. on Present-Day Problems of Power Engineering (APE ‘01), Gdansk–Jurata, June 6–8, vol. III, pp. 75–82Google Scholar
  7. 7.
    Matusiak B. (2000) Hybrid Neural — Genetic System for Energy Transaction Planning. Proc. of the III Conf. on Theory and Applications of Artificial Intelligence (CAI 2000 ), Lódi, Poland, October 5–7, 41–48Google Scholar
  8. 8.
    Matusiak B., Bartkiewicz W. (2001) Linking neural predictors with decision models: The short-term load forecasting case study. Proc of X-th Intl Conf. on Present-Day Problems of Power Engineering (APE ‘2001), Gdansk–Jurata, Poland, June 6–8, vol. III, 109–116Google Scholar
  9. 9.
    Rules of the Balancing Market 2000 in Poland ver. 1.0., PSE S.A., Warsaw, (in Polish)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2003

Authors and Affiliations

  • Witold Bartkiewicz
    • 1
  • Bożena Matusiak
    • 1
  1. 1.Department of Computer ScienceUniversity of ŁódźŁódźPoland

Personalised recommendations