基于自适应PSO-BP神经网络的电力工程造价预测研究  

Research on Power Engineering Cost Prediction Based on AdaptivePSO-BP Neural Network

作　　者：于炳慧 Yu Binghui(Foshan Tiancheng Engineering Consulting Management Co.,Ltd.,Foshan 528000,China)

机构地区：[1]佛山市天诚工程咨询管理有限公司,佛山528000

出　　处：《办公自动化》2024年第21期1-3,共3页Office Informatization

摘　　要：文章提出对基于自适应PSO-BP神经网络的电力工程造价预测方法的设计与研究。根据当前的预测需求,先进行数据预处理,用多阶段的形式,扩大预测的覆盖范围,完成设定多阶段造价预测的目标。基于此,设计自适应PSO-BP神经网络电力造价预测模型,用动态寻优的方式实现最终预测处理。测试结果表明:对比于大数据电力工程造价的预测方法、GIM标准电力工程造价预测方法,文章设计的自适应PSO-BP神经网络电力工程造价预测方法最终得出的平均误差相对较小,整体上较可控,这说明在自适应PSO-BP神经网络的辅助下,文章设计的电力工程造价预测方法更加高效、稳定,针对性明显提升,造价预测的效果更为真实。The design and research of a power engineering cost prediction method based on an adaptive PSO-BP neu-ral network is proposed.According to the current forecast demand,data preprocessing is conducted first,and a multi-stage approach is adopted to expand the coverage of the forecast and complete the setting of stage cost forecast tar-gets.Based on this,an adaptive PSO-BP neural network power cost prediction model is designed,and a dynamic opti-mization method is used to achieve the final prediction.Test results show that,compared with the big data power engi-neering cost prediction method and GIM standard power engineering cost prediction method,the average error of the proposed adaptive PSO-BP neural network-based method is relatively small and more controllable overall.This indi-cates that,with the assistance of the adaptive PSO-BP neural network,the designed power engineering cost prediction method is more efficient,stable,and targeted,with a more accurate cost prediction effect.

关 键 词：自适应结构 粒子群优化-反向传播(PSO-BP)神经网络 电力工程 造价预测 成本控制 电力系统 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TU723.3[自动化与计算机技术—计算机科学与技术] TM7[建筑科学—建筑技术科学]