基于ABC—BP神经网络的用电量预测研究  

Research on Electricity Demand Forecasting Based on ABC—BP Neural Network

作　　者：徐晨[1] 曹莉[2] 梁小晓[3] 乐英高[2] 

机构地区：[1]三峡大学电气与新能源学院,湖北宜昌443000 [2]四川理工学院自动化与电子信息学院,四川自贡643000 [3]四川理工学院计算机学院,四川自贡643000

出　　处：《计算机测量与控制》2014年第3期912-914,922,共4页Computer Measurement &Control

基　　金：四川省高校重点实验室项目(2013WYJ03)

摘　　要：针对区域用电量的时效性、复杂性和非线性等特点,提出基于人工蜂群算法(ABC)优化BP神经网络(ABC-BP)的区域用电量预测分析模型;以BP神经网络为基础,将往年区域用电量作为用电置的预测样本,采用基于ABC算法对BP神经网络的各个权值和阈值参数进行优化,最后建立模型应用于区域用电量预测系统,为分析区域内经济发展水平、经济走势、产业分布状况及政策实施效果等问题提供有力支持;介绍了人工蜂群算法(ABC)和BP神经网络算法,详细阐述ABC算法优化BP神经网络的权值和阈值;通过实验仿真对比,提出的算法预测结果比仅仅使用BP神经网络算法以及粒子群优化BP神经网络算法更高,是一种有效可靠的区域用电量预测方法。In view of the regional consumption of electricity efficiency, complexity and nonlinear characteristics, proposed the regional consumption prediction model based on artificial bee colony algorithm （ABC） optimizes BP neurotic network algorithm （ABC--BP）. With the BP neural network as the foundation, the usual area power consumption as the electricity consumption forecast of sample, using ABC-- BP neural network algorithm to optimize the various parameters of weights and threshold, finally establish the model put into use the regional electricity consumption forecast model, for analyzing the level of economic development in the region, economic trends, industry distribution, and provide strong support to policy implementation and other issues. This article describes the ABC algorithm and BP neural network algo- rithm, elaborated on the ABC algorithm to optimize BP neural network weights and threshold. Through the comparison of the experimental simulation, forecasting results obtained by this method better than BP neural network algorithm, get the higher accuracy, and it is an effec- tive and reliable method.

关 键 词：人工蜂群算法 BP神经网络 用电量预测 预测算法 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]