Prediction Of EPC Cost Index Based On PSO-ACO-BP Neural Network

EPC engineering general contracting is widely favored by the market as a construction engineering organization and implementation method that covers the entire life cycle of project construction.In the application of EPC project general contracting mode,the majority of investment forms are state or collective investment.Therefore,the cost control and management of the project have an important connection with the rational use of state-owned assets.The current project valuation and measurement system in China is mainly based on the construction drawings,which can determine a more accurate project quantity and comprehensive unit price.For projects adopting the EPC project general contracting model,the construction unit can use the general contracting method to issue the contract after the construction project is approved and approved,or after the scheme design is approved,or after the preliminary design is approved.At this time,the detailed construction drawings of the project are not designed,which leads to the inability to determine the project’s accurate engineering volume and the comprehensive unit price corresponding to the detailed design standards.The project cost cannot be accurately determined,which is not conducive to the contractor’s control of the project investment and is not conducive to contracting.Human optimization design and improved construction have restricted the implementation of the project general contracting model.According to "Jianban Biaohan [2018] No.726 of the General Office of the Ministry of Housing and Urban-Rural Development,"Specifications for Measurement of General Contracting Pricing of Housing Construction and Municipal Infrastructure Projects"(Draft for Comment),the determination of construction and installation costs during the contract issuance period can be determined Refer to this cost of similar or similar projects and consider the price index.In view of the fact that China has not made relevant research on the cost index under the EPC general contracting model,it is very important to establish an EPC general contracting cost index system and compile and forecast the cost index based on the relevant documents of the EPC general contracting project necessary.This thesis is based on the theory of engineering cost index,according to the actual needs of the participants of the EPC general contracting project,and in conjunction with the "Construction Project Total Investment Cost Project Composition" and "Construction Project General Contracting Cost Project Composition" issued by the Ministry of Housing and Urban-Rural Development for comments "Draft" Jianban Biaohan [2017] No.621,established the cost index system of EPC general contracting projects,in order to fill the gap of the domestic EPC cost index system;drawing on the domestic and foreign research on the cost index and the cost index prediction model,Decided to use the BP neural network model with strong data learning ability as the basis,and take the EPC general contracting project as the research object to establish an EPC cost index prediction model based on the PSO-ACO-BP neural network,and apply the model to the EPC general contracting project Forecast of construction and installation engineering costs.The main research contents of the paper are:The basic theory of engineering cost index is expounded,and the EPC cost index system suitable for the whole process of design,procurement and construction is established.According to the cost structure of the general contracting project,the EPC general contracting cost index system and the indexes that need to be researched are preliminarily designed,and the cost index in the construction and installation cost is selected as the research object of this article.Several commonly used engineering cost index prediction models are studied and analyzed,and it is determined to use the BP neural network prediction model as the research model in this paper.The basic theory of traditional BP neural network is explained,the advantages and disadvantages of BP neural network are analyzed,and its shortcomings are improved.Finally,the EPC cost index prediction model is constructed,namely: PSO-ACO-BP neural network prediction model.The actual data of the EPC general contracting project were substituted into the BP neural network prediction model and the PSO-ACO-BP neural network prediction model,and the two models were used to train and predict the construction cost index respectively,and the predicted value of the EPC construction index was obtained.The analysis of the prediction results of the two models proves that the PSO-ACO-BP neural network prediction model adopted in this paper has higher prediction accuracy than the traditional BP neural network prediction model.Finally,the model is applied to the determination of the construction and installation cost of the EPC general contracting project,so as to verify the feasibility and effectiveness of the prediction model.The research results of the paper on the EPC cost index can be used to guide the project estimate and determine the budget during the EPC general contract project issuance and contracting stage,and assist the contractor in the project bidding stage to judge whether the bid price is reasonable.In the implementation phase of the project,the prediction of the cost index can enable the project manager to have a certain control over the future cost change trend,which provides a reference basis for determining the construction and installation engineering cost of the project;in the promotion of the EPC general contracting model,it is Increasing capital utilization provides strong support.