| 王江荣.基于卡尔曼滤波和粒子群优化算法的灰色神经网络预测模型[J].电气自动化,2014,36(1):24~26 |
| 基于卡尔曼滤波和粒子群优化算法的灰色神经网络预测模型 |
| Gray Neural Network Prediction Model Based on Kalman Filtering and Particle Swarm |
| |
| DOI： |
| 中文关键词: 卡尔曼滤波 G(1,1)模型 预测 PSO算法 BP神经网络 |
| 英文关键词:Kalman filtering G (1,1) model prediction PSO algorithm BP neural network |
| 基金项目: |
|
| 摘要点击次数: 2725 |
| 全文下载次数: 0 |
| 中文摘要: |
| 针对波动大且具有非平稳性的负荷序列预测问题，建立了基于卡尔曼滤波和粒子群优化算法的灰色神经网络预测模型（R.E Kalman-G(1,1)-PSO-BP）。利用了卡尔曼滤波算法能够剔除非平稳序列中的随机误差，以获得逼近真实情况的有效信息的特点，对负荷测量序列进行滤波处理，根据GM(1,1)模型算法对滤波后的量测序列进行拟合预测。利用基于粒子群优化算法的BP神经网络算法对残差进行修正，得到了新的预测值。实践表明新预测值的整体精确度远高于GM(1,1)模型及Kalman-G(1,1)模型的预测精度。因此，所建模型具有较高的使用价值。 |
| 英文摘要: |
| With respect to the prediction of volatile and non stationary load sequence, this paper established a gray neural network prediction model based on the Kalman filtering and particle swarm optimization algorithm (R.E Kalman G (1,1) PSO BP). The model uses the Kalman filtering algorithm to eliminate random errors from the non stationary sequence and obtain useful information close to the real situation, so that the load measurement sequence is filtered. The measurement sequence after such filtering undergoes fitting prediction according to the GM (1,1) model algorithm. The residual error is corrected in the BP neural network algorithm based on the particle swarm optimization to obtain a new predictive value. Practice shows that the overall accuracy of the new predictive value is much higher than that of GM (1,1) model and Kalman G (1,1) model. Thus, the model established in this paper has a high value in use. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
