Research On Investment Estimation Method Of Jacking Frame Bridge Based On Improved BP Neural Network

With the continuous development of China,China’s railroad and highway mileage is increasing at the same time,the new planning line and the existing railroad or highway crossings are also increasing.In order to reduce the impact on the existing lines and for the safe operation of the existing lines,the overhead frame bridges are often used in the crossover leveling project.At the same time,due to the rise of artificial intelligence technology,BP neural network is applied to all aspects of engineering management.At present,there are complex and not easy to identify the influence factors in the jacking frame bridge project estimation,the estimation index is not fully applicable,and the deviation of the estimation results caused by cross-discipline and multi-discipline is large.In this context,this paper investigates the investment estimation method for jack-in frame bridges from the jack-in frame bridge project itself and reinforces the use of historical data from previous projects.Firstly,the construction process of jacking frame bridge and the composition and requirements of jacking frame bridge investment estimation and the existing problems are analyzed.Then the basic theory and process of the method used in this paper are explained.Then,by summarizing the applicability and shortcomings of various investment estimation methods in previous studies,the investment estimation model method of improved BP neural network is proposed,and the theoretical basis of genetic algorithm to improve BP neural network is introduced.Secondly,the principle and process of determining the impact factor of jacking frame bridge are introduced.the project decomposition and principal component analysis are used to narrow the scope.Then,the literature is used to extract the impact factor and the questionnaire survey method,and the weight of each expert score is given according to the expert background score,and then the impact factor is determined by comprehensively calculating the score of each impact factor.Thirdly,the topological structure of the model and the number of nodes in each layer are determined,and the method and process of improving BP neural network by genetic algorithm are determined.Then,the estimation model of improved BP neural network is established,and the model is realized by Matlab platform.At the same time,some graphical user interfaces of the system are completed.Fourth,the data of previous projects are input into the model established in this paper as samples,and the training group and test group are divided for verification.By tracking the operation process and results of the model,the index values are calculated to test the validity of the model.Finally,through the output results of the divided test group,it is found that the model in this paper works better than the BP neural network before improvement,and the output results of the model in this paper are closer to the real value and more stable,which can meet the accuracy requirements of the investment estimation of the jacking frame bridge,and provide new ideas and methods for the investment estimation work.