Study On Cost Estimation Of Hydro-electirc Engineeirng Project Based On Bp Neural Network With Fuzzy Optimization

Project cost is an important criteria for project development as well as animportant task at the beginning of a project. At the early decision-making stage,because of lacking detailed engineering material, it is very difficult to make use oftraditional methods for cost estimation. Therefore, a new cost estimation methodapplicable to project feasibility study stage and conceptual design stage will havegreat significance.The combination of fuzzy mathematics and neural network is firstly used toestablish hydro-electric project cost estimation model. The fuzzy mathematics theoryis used for cost data filtering, aiming to select similar engineering material as thetraining samples. The BP neural network is chosen to realize the complicatednonlinear mapping between hydropower engineering characteristics and cost index.Finally, the well established model is used for new case’cost estimation. Throughcomparison of forecast value and actual value, the model is proved to be effectiveand practical.(1) Comments about investment development status and cost managementsituation of hydro-electric project is made, especially those related literature ofestablished estimation model based on fuzzy mathematics and neural networkscarefully reviewed.(2) Based on two major theoretical foundations, hydro-electric project costestimation model is set up. with implementation of PBS decomposition ofhydro-electric project, comparison layer projects are got and well established; Asinterpretative structural modeling theory used to analyze cost impacting factor,feature category index systems of comparison layer projects are finally set up, whichpaves the way for fuzzy optimization and cost estimates.(3) How cases fuzzy optimization of hydro-electric project carry out. Based onfuzzy mathematical theory, combined with engineering reality of comparison layerprojects, membership function and weights of feature category index system are got ,with approaching degree calculated(4) BP neural network can be applied to hydro-electric project cost estimation. Based on BP neural network theory, hydro-electric project cost estimation model isestablished, including each sub-model parameters design as well as input andoutput data preprocessing. Cost estimation model can be achieved in Matlab.(5) Case study. According to the established cost estimation model, ahydro-electric project cost is estimated and a comparison between forecast valueand actual value is conducted. The achieved result is satisfied, which verifies thevalidity and practicality of cost estimation model.This thesis is only related to preliminary research on model establishmentand cost estimation method. Sample database established effectively, modelgeneralization ability improvement and visualization operation platformdevelopment are expected to be further researched.