基于粒子群优化算法的中国能源需求预测

Energy Demand Forecast in China Based on Particle Swarm Optimization Algorithm

能源需求预测是能源规划和政策制定的基础。经济增长、总人口、产业结构、城市化率、能源消费结构以及技术进步等因素都会影响能源需求。采用粒子群优化算法,通过两种函数形式(线性和指数)建立了基于影响因素的能源需求预测模型。以1980-2005年的各指标数据作为训练样本进行模型的参数估计,并使用2006-2010年的相关数据作为检测样本来验证所建模型的有效性。模拟结果显示,指数模型的预测能力略优于线性模型,但两者的能源消费量预测值都接近真实值,估计误差较小:拟合部分的平均相对误差分别为0.76%和0.57%,检测部分的平均相对误差分别为0.78%和0.624%。这说明粒子群优化算法在解决我国能源系统非线性及高维识别问题上具有一定的有效性。最后,通过分析各影响因素的变化趋势,对中国2011-2015年的能源需求进行了预测。在拟定各因素的增长率下,我国能源消费量从2011年的343 668.5万t标煤上升到2015年的432 169.5万t标煤,年均增长率为5.9%。此结果表明,十二五期间我国能源需求形势依然严峻。

Energy demand forecast is fundamental to energy planning and policies formulation. There are many factors such as economic growth, population, industrial structure, proportion of urban population, energy consumption structure and technological progress that can affect energy demand. Through particle swarm optimization algorithm, two forms of equations （linear and exponential） are proposed to establish energy demand forecast model based on affecting factors. The training sample between 1980 and 2005 is applied to estimate the coefficients of the model, and the testing sample between 2006 and 2010 is used to verify the established model. The results show that the predictive ability of exponential model is slightly better than the linear one, but the predicted values of both are close to actual values with a smaller estimation error： the average relative errors are 0.76% and 0.57% in fitting part, and are 0.78% and O. 624% in predicting part, respectively. This turns out to be effective for particle swarm optimization algorithm to solve nonlinear and high dimensional identification problem of China energy system. Finally, energy demand over the years 2011- 2015 are predicted by analyzing the trends of affecting factors. Based on the growth rate of affecting factors set in this paper, China＇ s energy demand will increase from 3 436. 685 millions of tons of standard coal （Mtce） to 4 321. 695 Mtce in 2015, with an average growth rate of 5.9%, which indicates a grim situation still exists in the 12th Five Year Plan period of China.