Research On Estimation Model Of Construction Cost Based On Optimal Neural Network

In the valuation process of the construction project cycle for many times,the most rough is the investment estimation,but it has a great impact on the whole project.It is the basis of project decision making and an effective tool to make investment plan and control investment.However,due to the early stage of the project cost estimation,a lot of information about the project can not be determined,and there will be many unforeseen events affecting the project cost in the process of project construction,making the preparation of project investment estimation very difficult.Therefore,it is of great significance to find a feasible method of rapid investment estimation for construction projects.First of all,this paper is based on a large number of reference and in-depth study of domestic and foreign construction project investment estimation methods and estimation models,the data of the 49 sets of residential project cost data collected and the proportion of the project cost to the total project cost are statistically analyzed and analyzed.According to the primary and secondary factor analysis method,the main sub-projects in the construction and installation project cost are found out.Extract and sort out the engineering characteristics.Then the hierarchical structure model is established to analyze the engineering feature vector that affects the project cost.According to the primary and secondary factor analysis method,the secondary engineering feature factors are eliminated.The main engineering feature vector selected by the screening is used as the input vector of the engineering cost estimation model,and the engineering characteristic factors are analyzed.The mechanism of influence on project cost is used as the input vector of the estimation model.Secondly,according to the learning process of standard BP neural network,combined with the relevant knowledge of gray number in grey theory,the interval gray value is used to set the initial connection weight between the hidden layer and the input layer,and it is continuously adjusted during the training process.The value interval of the interval gray number approximates the optimal value of the weight,which improves the learning process of the standard BP neural network.Thirdly,in order to verify the validity and feasibility of the model,in the early stage of the project,according to the 117 samples of the collected residential projects,,and the remaining 1 was test samples,and the engineering feature vector was quantified.After the gray accumulation and normalization of the training samples and test samples,the MATLAB software was used to establish a dynamic and rapid estimation of the construction cost model.The prediction results of the gray BP neural network model met the requirements.Through specific engineering examples,the GM(1,1)model,BP neural network method and grey BP neural network method in grey system theory are used to predict respectively,and the BP neural network after gray theory optimization is obtained.Practical conclusions of the network model in the application of the construction industry.Finally,taking the 2#building of a residential district in Benxi as the empirical analysis object,using the established model and collecting similar engineering data training model,and estimating the relevant cost data of 2#building,and using the estimation index to calculate 2#The unilateral cost of the building and other data.By comparing the BP neural network estimation model prediction value and the error between the traditional estimation method and the actual value,the accuracy,efficiency and reliability of the established model are verified.This paper intends to establish a construction engineering cost estimation model based on grey BP neural network.This model provides technical support for the estimation of construction project cost to effectively reduce engineering cost and improve work efficiency.The construction project cost estimation is in the direction of automation and intelligence.Development has certain practical significance.