For the international competition "GECCO 2012 Industrial Challenge" computer scientists developed various methods to significantly improve the prediction of future power consumption based on smart meter data. The competition was organized by the Cologne University of Applied Sciences and the software provider GreenPocket.
As part of the "Genetic and Evolutionary Computation Conference (GECCO)" the Industrial Challenge took place from 7 to 11 July 2012 in Philadelphia (USA). The goal of this year's event was to develop accurate forecasting methods for electrical energy consumption profiles.
Participants from all over the world were provided with real energy data of three months. The task was to predict the consumption for the following month as accurately as possible. Stephan Hutterer, a researcher at the University of Applied Sciences of Upper Austria, developed the winning algorithm.
Altogether Professor Thomas Bartz-Beielstein (Cologne University of Applied Sciences) was very satisfied with the outcome of the event: "We have seen absolutely convincing approaches that improve the reliability of forecasts significantly and provide an added value for the further development of smart metering."
To create an authentic basis for the competition, the Cologne University of Applied Sciences collaborated with GreenPocket, a leading provider of smart meter and smart home software. The company provided the smart meter consumption data and its own forecast calculations.
"The precise forecast of electricity consumption is very important for the quality of our products. Therefore, we are pleased when top-class scientists deal with this topic", says Dr. Thomas Goette, CEO of Green Pocket.
