Affiliations: [a] School of Economics and Management, Harbin Engineering University, Harbin, China
| [b] Department of Physics, University of Science, Pyongyang, Korea
| [c] Department of Physics, Kim Chaek University of Technology, Pyongyang, Korea
Correspondence:
[]
Corresponding author. Zhilong Song, School of Economics and Management, Harbin Engineering University, Harbin, 150001, China. Tel.: +8613019703877; E-mail: [email protected].
Abstract: The ability to accurately and reliably predict annual electricity demand is essential in modern society for effective planning, economic development, and to ensure the sustainability of the electricity supply. Considering the correlation between annual electricity demand and economic development, as well as annual electricity demand under low-carbon-economy targets, this paper proposes an improved quantum clustering algorithm (particle swarm optimization–weighted distance quantum clustering, PSO-WDQC) as a power demand forecasting model. This method can not only improve the accuracy of predictions but also accurately evaluate the economic development of a region. To demonstrate this ability, the paper applies the proposed method to low-dimensional Iris data as well as high-dimensional Wine data in order to verify the effectiveness of the method. Then, the method is combined with ridge regression to predict the demand for electricity under the low-carbon-economy target of China. The experimental results show that the method can accurately predict annual power demand with a relative error of 0.1674%. Moreover, the model accurately reflects that the Chinese economy has entered a new normal state since 2012, meaning that the economic growth rate has changed from high-speed to medium-high-speed.
Keywords: Particle swarm optimization, weighted distance, quantum cluster, electric power demand, prediction
