投资项目WLCS的Hopfield预测模型研究

Research on the Estimating Whole Life Cost-significant of Construction Projects Based on the Hopfield Neural Network

研究目的:当前我国土木工程基础设施正处于大规模建设时期,但相应的投资估算却出现一些不相适应的现象,这与传统的造价估算模式不无关系,传统的造价估算模式大多采用线性方法,较为简单直接地对造价和估算方法建立模型,但实际中项目造价因存在种种影响因素和差异,不可能归为简单线性的性质。因此本文将全生命周期显著性造价理论与人工神经网络相结合,研究和探索非线性工程造价估算模型的可建立性和准确性。研究结论:(1)在全生命周期造价中采用显著性造价方法,能够快速准确的找出实际数据,大大提高效率和精度;(2)将上述方法与传统预测方法进行对比,可以发现运用Hopfield神经网络提高了计算精度,简化了计算过程,减少了计算时间,Hopfield预测模型能够快速实现工程显著性项目的联想记忆,从而可以准确估算拟建工程的预算造价;(3)本文提出的投资项目估算方法和评价体系,可应用于基础建设项目领域的建设项目成本估算和控制等工作。

Research purposes： The transport infrastructure is in a large - scale, high - standard construction period at present in China. However, the investment management is incompatible with the increasing of government investment. The main reason for these problems is the existing cost estimation model which mostly uses linear method, and the actual project cost cant be classified as simple linear properties due to various factors and differences. This paper combined the whole life cost- significant theory with artificial neural networks, studied and explored the realizability and accuracy of nonlinear methods for investment project cost. Research conclusions： （ 1 ） Using cost - significant method in the whole life cycle cost can quickly accurately find out the actual data, greatly improve the efficiency and accuracy. （2） Through contrasting the prediction precision of WLCS and CS, the method of Hopfield neural network can improve the calculation precision, simplify the calculation process, and reduce the computation time. Hopfield model can quickly realize the associative memory of significant project, and accurately estimate the budget of the proposed construction project cost. （3） The estimation method and evaluation system can be applied to cost estimation and control in the field of infrastructure projects.