Authors: Zaid Mohamed, Pat S. Bodger
Addresses: Department of Electrical and Computer Engineering, University of Canterbury, PO Box 4800, Christchurch, New Zealand. ' Department of Electrical and Computer Engineering, University of Canterbury, PO Box 4800, Christchurch, New Zealand
Abstract: The logistic model has been very effective in forecasting many technological forecasting patterns. However, it has the characteristic of underestimating the forecasts in many situations. This is mainly due to the constraints imposed by the constant saturation level of the logistic growth curve. This paper proposes a variable asymptote logistic (VAL) model for forecasting electricity consumption in New Zealand. The saturation level of electricity consumption is found by the respective degree of correlation with the population of the country and the price of electricity. This is used as a variable asymptote in this logistic model and hence the name variable asymptote logistic (VAL) model. The VAL model gave lower forecasting errors than the logistic model and gave future forecasts that are higher than the logistic model. The software package MATLAB is used in all stages of the simulation.
Keywords: forecasting electricity; forecasting accuracy; variable asymptote logistic model; electricity consumption; energy consumption; New Zealand; saturation level; forecasting errors; simulation; MATLAB.
International Journal of Computer Applications in Technology, 2005 Vol.22 No.2/3, pp.65 - 72
Published online: 26 Apr 2005 *Full-text access for editors Access for subscribers Purchase this article Comment on this article