摘要
针对装修工程造价计算流程复杂、难以快速估算的问题,结合BP神经网络构建面向简单装修工程的工程造价估算模型,以15个典型工程特征作为输入指标,实现对工程造价和人工工日的快速估算。利用Keras与Tensorflow编写程序,选取典型工程样本训练模型,并进行了验证分析。结果表明,该模型能够较好地学习输入与输出之间的关系,有效预测工程造价与人工工日,为装修工程造价的快速估算提供一种有效途径。
The complicated calculation process and difficult quick estimation are the main problems in decoration project cost. Combined with BP neural network, an engineering cost estimation model for simple decoration project is constructed,which takes 15 typical engineering characteristics as input indicators to realize rapid estimation of engineering cost and manual work cycle. Keras and Tensorflow are used to write the program, and the typical engineering sample training model is selected and verified. The results show that the model can well establish the relationship between input and output, effectively predict the project cost and manual working time, which provides an effective way for the rapid estimation of decoration project cost.
引文
[1]陈菁瑶,苗鸿宾,刘兴芳,等.基于BP神经网络的钻削力预测研究[J].机械设计与研究,2018,34(3):116-118.
[2]丁硕,巫庆辉.基于改进BP神经网络的函数逼近性能对比研究[J].计算机与现代化,2012(11):10-13,17.
[3]张敬玲.BP神经网络的应用[J].石家庄职业技术学院学报,2015,27(4):34-36.
[4]任宏,周其明.神经网络在工程造价和主要工程量快速估算中的应用研究[J].土木工程学报,2005(8):135-138.
[5]孟俊娜,梁岩,房宁.基于BP神经网络的民用建筑工程造价估算方法研究[J].建筑经济,2015,36(9):64-68.
[6]张登文,蒋红妍,张子圆.基于BP神经网络的建筑工程造价快速预测[J].水利与建筑工程学报,2010,8(3):61-62.
[7]梁喜,刘雨.基于模糊神经网络的建筑工程造价预测模型[J].技术经济,2017,36(3):109-113.
[8] KINGMA D P,BA J.Adam:a method for stochastic optimization[EB/OL](2017-01-30)[2018-07-05].https://arxiv.org/pdf/1412.6980.pdf.