Research On The Intelligent Method Of Cost Estimation In High-speed Railway Project

Investment estimation in the feasibility study plays a crucial role in project decision and DB,EPC project management models.The investment estimation of the HSR project is an important part in the plan,proposal and feasibility study report.It is also an important basis of engineering investment control.However,the traditional cost estimation method is single,linear and time-lag.With AI developing fast,based on the study of the intelligent HSR construction investment estimation,the simple way of estimation was promoted.Intelligent estimation algorithms were applied to the article to build investment estimation models which were set on different characteristics of the project to achieve dynamic and non-linear investment control.It is definitely important to control the HSR construction investment.Firstly,on the basis of significant cost theory,the methods of literature collection,grey correlation analysis and expert interviews were used to qualitatively identify,analyze and verify the significant cost factors.Secondly,cosine similarity was used in the database of established completed projects to select high similar projects of the proposed project,and then initially established intelligent investment estimate models.Finally,based on the similarity recognition,MATLAB R2016 a software was used to make intelligent estimation.If there was a large amount of completed engineering data,the BP neural network was used for cost prediction.If the number of completed similar projects was small and the similarity was not as high as the latter,and the BP neural network had not used for convergence,the fuzzy C-means was used.The completed and proposed projects were clustered,and then the cost of proposed project was identified by the traditional calculation formula.The experimental results showed that the error between the intelligent estimation algorithm and the budget value was within ±5%,so the intelligent models were feasible.In addition,if the existing available information quantity of the proposed project was significantly less than the selected significant cost factors,the significant cost factors were firstly clustered,then input clustered cost factors and fuzzy rules which affected HSR investment estima-tion to the fuzzy inference system,and the ratio of the proposed project to the completed project cost was obtained.Then the cost range of the proposed project could be obtained.