摘要
由于煤电价格的波动受多种不确定因素的影响,且煤价和电价之间还存在非常复杂的耦合关系,它是一个典型的非线性系统,所以使用传统的方法来建立煤电价格的预测模型非常困难。针对这种情况,提出了一种基于粒群算法(PSO)和BP神经网络的煤电价格预测方法。采用PSO训练BP神经网络,不仅克服了BP神经网络算法易于陷入局部最优的缺点,而且可以提高网络的收敛速度和预测精度。结合煤电价格的历史数据,在Matlab平台上进行了仿真实验,验证了该预测模型的优越性。
Since the price of coal and electricity depends on various indeterminate factors, there is a very complicated coupling relationship between the price of coal and electricity and this prediction model is a typical nonlinear system, it is hardly possible to set up a precise prediction model with the traditional methods. This paper proposes a means, on the basis of PSO ( particle swarm optimization) and BP neural network, to predict the price of the coal and electricity. With the introduction d PSO and BP neural network, this method not only avoids the shortcoming of the BP-neural-network way, which is susceptible to coming up with a partial- optimal result. Moreover, it can improve the constringency speed and the predicting exactitude of the network. Combined with the historical statistics of the coal and electricity, we carry out an emulation experiment on the platform of Matlab, thus verifying the superiority of this predicting model.
出处
《后勤工程学院学报》
2006年第3期92-95,101,共5页
Journal of Logistical Engineering University
关键词
BP神经网络
粒群算法
煤电联动
BP neural network
PSO
linkage motivation d the coal and electricity