摘要
提出了一种 BP混合模拟退火 (SA)的 ANN短期负荷预测方法 ,该方法针对传统 BP学习算法的缺点 ,将 BP算法和模拟退火算法的优点相结合以提高网络的学习性能。 ANN模型中考虑了温度和预测日类型 ,可进行工作日和节假日的预测 ,实例表明 ANN模型实用有效、精度高。
This paper proposes a neural network (ANN)with the hybrid learning strategy which combines the back propagation(BP) with simulated annealing (SA) algorithm. As back propagation learning algorithm has some drawbacks, BP & SA hybrid learning algorithm, which combines the property of BP with the property of SA algorithm, is adopted to improve the learning property. The effects of temperature and day of the week are considered in ANN model. Loads of weekdays and holidays can be forecasted based on the ANN model. Numerical tests showed the high efficiency and accuracy of the ANN model.
出处
《合肥工业大学学报（自然科学版）》
CAS
CSCD
2000年第1期78-80,共3页
Journal of Hefei University of Technology：Natural Science
基金
国家攀登 B计划资助!项目 (930 2 110 10 )
安徽省自然科学基金!资助项目 (974 130 0 1)
