Engineering Cost Estimation Research Based On Improved BP Neural Network

With the rapid development of China’s construction market bidding system,step-by-step implementation of the contract system, as well as accession to the WorldTrade Organization (WTO) with the international practice needs, Engineering CostEstimation based on the deepening of reform,"construction project BOQ pricingnorms"The introduction of rapid valuation of construction enterprises put forwardhigher requirements. In recent years, the increasingly fierce competition in the marketof construction bidding, tender offer period of time is getting shorter and shorter, fastand accurate project cost estimates, construction enterprises is essential.Use the Analytic Hierarchy Process (AHP) to analyze the factors affecting project costto determine the characteristic factors of the project, and the domain of its value, isdivided into two sub-space of the exact amount and the amount of blur, fuzzy theory offuzzy characteristic elements for the solution of the fuzzy membership; advantage of theimproved BP neural network to establish the project cost estimation model, in order todetermine the impact of factors as input variables, the unilateral cost as the outputvariable, thereby establishing important factor with unilateral cost between nonlinearfunction of sample training, access to the network approximation ability to achieve aquick estimate of the project cost.In this paper, written in MATLAB software project cost estimation procedures; thepredicted unilateral cost estimates value the correction and adjustment, and combinedwith the the Vague set matching theory to control the total project cost; engineeringexample to verify the feasibility of the model provide a theoretical basis for theconstruction enterprises Quick Quote. Thus, the organic combination of improved BPneural network estimation model and Vague set matching theory, meet the sameaccuracy, training speed, error is small, the predicted value and the actual value of thetotal cost of a building project, get satisfactory results.