Research On The Cost Estimation Of Construction Projects

With the continuously steady and rapid economy development of China, the construction industry outstanding contribution to national economic growth recent years. Faced with the complexity of the current country’s macroeconomic environment and the increasingly fierce competitive situation, it has become more and more important to estimate the investment of projects quickly and accurately. However, the currently widely used methods requires not only accurate engineering drawings and engineering quantities, such as standard and physical method, which are time-consuming and labor-consuming, but also affected by the staffing’s levels and subjective factors, so that make it difficult to estimate the budget cost. This paper established the corresponding prediction model according to the neural networks theory and case-based reasoning theory, thereafter the analysis based on the principal component analysis were studied to furtherly improve the accuracy of the project cost pridictions. Therefore, the research has a certain theoretical and practical significance.The main works are as follows:1. This paper analyzes and summarizes the status of the background in domestic and abroad, elaborated the existing main forecasting model estimation methods and corresponding estimates in the field of project cost estimation. Furthermore, an emphasis on case-based reasoning models were introduced, including the basic theory of case-based reasoning, the implementation steps and the characteristics. The paper established the framework of the prediction model based on case-based reasoning, which laid the theoretical foundation for the future study.2. Established BPNN cost estimation model, first introduced the basic theory of neural networks, including the basic idea of the neural networks and it’s algorithm processes. Then establish the historical cases libraries, build neural network model for case-based reasoning, and validate the neural network and make prediction. At last, the average error percentage (Mean Average Percent Error, MAPE) of forecasting accuracy are analyzed and compared.3. Established the CBR-GA cost estimation model. Firstly, introduced the basic theory of genetic algorithms, mainly including basic idea, the implementation process and the characteristics. Then based on the historical cases libraries, this paper established the case similarity model used for case-based reasoning and use the genetic algorithms optimization method to solve the attribute weights. Finally, forecast the project costs based on the models, use the MAPE to evaluate the prediction accuracy are analyzed and compared.4. Combined with the theory of principal component analysi, this paper makes improvements for the proposed construction cost estimates model, established the PCA-BPNN model and PCA-CBR-GA model, including principal component analysis theory, using principal component analysis to filter the main attribute indicators, established the revised model on the basis of the BPNN model and CBR-GA model, then validate the data again and use the MAPE indicators to evaluate the predictive accuracy, finaly compared BPNN with the CBR-GA model.Research papers have a reference not only on the theory of intelligent algorithms to estimate the investment cost, but also on promoting the project estimates prediction in investment estimation practices.