基于小波神经网络的宏观经济对工程造价影响研究

发布时间：2018-05-30 14:14

本文选题：工程造价 + 宏观经济　；参考：《天津大学》2014年硕士论文

【摘要】：建设工程的造价是项目管理的一个重要方面,是投资者、业主、承包商、分包商以及其它利益相关者的内在驱动因素。我国虽然处于市场经济阶段,但是定额计价模式的观念仍然存在,如何在竞争激烈的国际舞台上立于不败之地,合理地预测建筑的造价显得至关重要。价格的预测仅从构成造价的人、材、机的变动来分析工程造价的变动已不能满足动态市场的要求,需要从宏观经济的角度来分析对工程造价产生影响的因素。只有充分把握宏观市场价格变动引起的工程造价的变动,合理预见风险,才能在竞标过程中得心应手。本文首先从理论角度分析了工程造价与宏观经济的相互关系,接着从六个方面定性地分析了影响工程造价的宏观经济变量:经济整体水平、物价变动水平、各行业发展水平、投资水平、生产率水平、建筑业竞争水平。在此基础上对小波分析、神经网络(ANN)及小波神经网络(WNN)进行了阐述并分析:得出小波神经网络不仅具有有效运用小波变换局部化性质的特点,还有神经网络自学习能力的特点,因而具备良好的逼近、容错能力以及收敛性能好、预测准确的特性。基于此从网络结构的确定、参数设置、算法等方面考虑建立了小波神经网络模型来研究宏观经济变量对工程造价的影响。利用输入层个数为1的小波神经网络模型确定影响工程造价的先导变量有国民生产总值、房屋施工面积、居民消费价格指数、建筑安装工程固定资产投资价格指数、按建筑业总产值计算的建筑企业劳动生产率、人均地区生产总值。将这些先导变量作为预测模型的输入变量,工程造价作为输出变量,验证了小波神经网络模型在工程造价预测方面有着较强的适用性。
[Abstract]:The cost of construction project is an important aspect of project management, which is the internal driving factor of investors, owners, contractors, subcontractors and other stakeholders. Although our country is in the stage of market economy, but the concept of quota pricing mode still exists, how to be invincible in the competitive international stage, it is very important to reasonably predict the construction cost. The forecast of price can not meet the requirements of the dynamic market only by analyzing the change of the construction cost from the change of the people, materials and machines that constitute the cost, and needs to analyze the influence factors on the project cost from the angle of macro economy. Only by fully grasping the change of engineering cost caused by the change of macro market price and reasonably anticipating the risk can we be able to handle the bidding process. This paper first analyzes the relationship between engineering cost and macro-economy from a theoretical point of view, and then qualitatively analyzes the macroeconomic variables that affect construction cost from six aspects: the overall level of economy, the level of price change, and the level of development of various industries. Investment level, productivity level, construction industry competition level. On this basis, wavelet analysis, neural network ANN) and wavelet neural network (WNN) are expounded and analyzed. It is concluded that wavelet neural network not only has the characteristic of using wavelet transform localization effectively, but also has the characteristics of neural network self-learning ability. Therefore, it has good approximation, fault-tolerant ability, good convergence performance and accurate prediction characteristics. Based on this, a wavelet neural network model is established to study the influence of macroeconomic variables on project cost from the aspects of network structure determination, parameter setting, algorithm and so on. Using the wavelet neural network model with input layer 1 to determine the leading variables that affect the project cost are gross national product (GNP), building construction area, consumer price index, fixed assets investment price index of construction and installation project. Construction enterprise labor productivity, per capita regional GDP calculated by the gross output value of the construction industry. These leading variables are used as input variables and project cost as output variables. It is verified that the wavelet neural network model has strong applicability in engineering cost prediction.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F124;TP183;TU723.3

【参考文献】