Summary
Proceedings of the 2012 International Symposium on Nonlinear Theory and its Applications
2012
Session Number:B1L-C
Session:
Number:320
Simultaneous prediction of temperature and demand for electric power by non-linear models with cross-correlations
Hiromasa Iwayama, Jun-Ichi Inoue,
pp.320-323
Publication Date:
Online ISSN:2188-5079
Summary:
We investigate the cross-correlation between temperature in Sapporo city and demand for electric power in our Hokkaido University. In order to predict the empirical time series, the so-called PUCK (Potential of Unbalanced Complex Kinetics) model, which is based on a potential-driven dynamics of Langevin-type, is modified in terms of information about cross-correlations. The modification enables us to predict the temperature and demand simultaneously. We will examine the statistical performance of the prediction for empirical data sets collected by Hokkaido University Office for a Sustainable Campus (http://www.osc.hokudai.ac.jp/) for every thirty minutes in August 2011. The resulting Temperature-Demand (TD) diagram is compared with the empirical evidence. Effect of the cross-correlation on the time series prediction and its usefulness are extensively discussed.
References:
[1] http://www.hokudai.ac.jp/
[2] http://www.osc.hokudai.ac.jp/
[3] http://www.data.jma.go.jp/obd/stats/etrn/index.php
[4] Y. Nameki, Proc. of the Operations Research Society Japan 1-A-7, (in Japanese) (1995).
[5] K. Watanabe, H. Takayasu and M. Takayasu, Physical Review E 80, 056110 (2009).