短期负荷预测的集成改进极端学习机方法  

Short-Term Load Forecasting Method Based on Ensemble Improved Extreme Learning Machine

作　　者：程松[1] 闫建伟[1] 赵登福[1] 王圈[2] 王海明[2] 

机构地区：[1]西安交通大学电气工程学院,西安710049 [2]河南省电力公司新乡供电公司,河南新乡453000

出　　处：《西安交通大学学报》2009年第2期106-110,共5页Journal of Xi'an Jiaotong University

基　　金：国家自然科学基金资助项目(50595411);陕西省攻关计划资助项目(2007K04-17);西安市科技创新支持计划资助项目(YF07040)

摘　　要：首次将极端学习机(ELM)理论引入短期负荷预测领域,并以BFGS拟牛顿法对ELM网络左侧权值进行优化训练调整,形成基于迭代-解析的改进ELM预测模型.同时,采用集成技术Boosting算法,生成多个差异度大的改进ELM子网络,对其进行加权组合,构建了集成改进极端学习机预测模型.该模型不仅有效避免了极端学习机左侧权值随机给定的输出稳定性问题,而且克服了单一网络预测模型泛化能力较差等缺陷.The extreme learning machine （ELM） theory is introduced into the domain of shortterm load forecasting, and an ELM forecasting model is developed based on iteration-parse via optimizing, training and rectifying the left weights with BFGS quasi-Newton method. Adopting boosting algorithm of ensemble technology, several quite different ELM child-networks are generated and weighted, and a forecasting model of ensemble improved ELM are constructed. This model avoids effectively the output unstability from the left weights given arbitrarily, and breaks such limitations as weak generalization capability of single network model.

关 键 词：极端学习机 短期负荷预测 训练 集成技术 

分 类 号：TM715[电气工程—电力系统及自动化]