Research On Compiling And Forecasting Of Comprehensive Cost Index Of Highway Engineering

The comprehensive cost index of highway engineering can provide reference for government departments to carry out macroscopic evaluation of highway industry,study the current situation of industry development and formulate highway construction development planning.The compilation of comprehensive cost index of highway engineering requires a large number of highway engineering cost data.Although the advent of the big data era,it is difficult to obtain enough data from reality due to the difficulty of data acquisition and high cost.The problem of"big data and small samples" is prominent.In this thesis,on the basis of reference to the domestic and foreign research on engineering cost index,Aiming at the disadvantages of poor applicability of construction cost index compiled by typical engineering method,great subjectivity and easily affected by the level of staff,etc.Starting from the source of data,a method of compiling comprehensive cost index of highway engineering based on small sample cost data using data statistics method is proposed.Firstly,the main characteristic factors affecting highway engineering cost are determined through literature research and related documents,and the collected highway engineering cost data are processed by using the Grubbs criterion to screen reasonable data.Then the highway engineering cost data generation model based on PCA-PSO-SVR is constructed,and the accuracy and stability of the data generation model are verified by MATLAB simulation analysis.Finally,the collected small sample data is used as the original data sample,and the PCA-PSO-SVR data generation model is used to generate a large number of highway engineering cost data samples that maintain the original data characteristics.Based on the generated highway engineering cost data samples,the data statistics method is used to prepare the comprehensive cost index of highway engineering,and the feasibility of the method is verified by comparative analysis.In order to predict the comprehensive cost index of highway engineering,this thesis considers the influence of annual highway construction mileage on highway engineering cost,and combines the functions and characteristics of GM(1,1)model,mind evolutionary algorithm and BP neural network to construct the GM(1,1)-MEA-BP prediction model for predicting the comprehensive cost index of highway engineering.The MATLAB software simulation analysis shows that compared with the single GM(1,1)model,the GM(1,1)-MEA-BP prediction model has higher accuracy.In this thesis,the highway engineering cost data generation model is constructed to expand the cost data sample,which provides a new idea for the acquisition of highway engineering cost data,and solves the problem that small sample cost data cannot be used to compile cost indicators by data statistics.Through the construction of prediction model,the comprehensive cost index of highway engineering is reasonably predicted,which has a certain role in promoting the construction of a complete highway engineering cost index system in the future,and has certain research value.