本文为加拿大康考迪亚大学（作者：Ines Siqueira）的硕士论文，共101页。
本文提出了一种基于神经网络的成本估算方法，主要为生成低层预制钢结构建筑的概念成本估算而开发。详细的成本估算是这类建筑的常用做法，因为成本估算往往受到各种不同参数的挑战。所开发的方法使用神经网络（NNs）对与项目直接成本相关的单个项目参数进行建模，它集成了成本调整的NN成本模型，允许以繁重工作量的方式评估不同的项目备选方案。NNs捕捉实际项目中遇到的问题，概括并利用这些知识来估算新项目的成本，使其成为应用程序中的一个非常强大的工具。本研究中使用的数据（75个建筑项目）是从加拿大一家大型预制钢结构建筑制造商（Canarn Manac）收集，数据持续时间为3个月。所开发的成本模型性能与未用于这些模型训练的项目所产生的成本以及通过回归预测的成本进行了比较。结果表明，当所提出的模型用于模型训练范围内的项目参数时，其性能优于回归模型。此外，所提出的模型可以解释并定义一个项目的许多参数，并对项目成本产生相当大的影响。所提出的方法可以很容易地进行调整，为风险管理提供决策支持，同时协助在广泛的行业中开发生产力模型。
This thesis presents a neural network-basedcost estimating method. developed for the generation of conceptual costestimates for low-rise prefabricated structural steel buildings. Detailed costestirnating is current practice for this type of buildings, since costestimators are often challenged by a wide variety of different parameters. Thedeveloped method ernploys neural networks (NNs) for modeling individual projectparameters associated with the direct cost of a project. It integrates NN costrnodels with cost adjustments. allowing for evaluation of different projectalternatives, in a tirnely manner. The ability of NNs to capture real lifeexperiences encountered on actual projects (Le. actual costs), generalize and utilizethat knowledge for estimating the cost of new projects makes it a very powerfultool to the application at hand. Data used in this study (75 building projects)were collected from a large manufacturer of prefabricated structural steelbuildings in Canada (Canarn Manac) over a 3-month period. The performance ofdeveloped cost models was tested against costs incurred by projects not used intraining of those rnodels, and costs predicted by regression. Results indicatethat the proposed rnodels, when used for projects with parameters within therange for which the models were trained, outperform regression. in addition,the proposed models can account for a number of parameters defining a project,and bearing considerable impact on the project cost. The proposed methodologycan easily be adapted to provide decision-support for risk management and toassist in developing productivity models in a wide range of industries.
- 引言
- 文献回顾
- 所提出的神经模型
- 本文开发的自动化成本估算系统
- 总结与结论
附录I 设计数据输入表
附录II 数据样本
更多精彩文章请关注公众号：