Establishment And Research Of Additional Quota For Highway Engineering Based On Artificial Neural Networks

With the rapid development of the technology, new processes, new equipment, new structures, new materials are constantly upgrading ,the existing highway project quota is often difficult to fully meet the requirements of road construction area, causing some projects are lack of realistic basis, now ti is urgent to establish additional quotas for each region.This paper describes the quota of domestic and international situation, and emphasis on the current quota of the preparation methods, proposed the importance and practical significance of the preparation of additional quota. Then Based the difference of quoto, Classify the factors to determine the main factors, using mathematical statistical methods to quantify them. Based on the study of the traditional method of highway additional quotas, to build a new model to calculate the data which can reflect the level of the production management,the aim is to lay a solid foundation for thesuitable establishment of the whole quota system.By introducing the Quota principle and the existed methods, analysis the current limitations of the method, subsequently, on the basic of analizing the characteristics of the original data to propose feasibility of establishing the forecast model of BP neural network for the preparation of quotas. Then introduced the BP neural network structure and algorithm theory, and application of BP neural network design for the original data model defined method and process, the establishment of BP network model and the model using Matlab programming language implementation. In the choice of training samples and testing samples, data on the impact of scale more detailed analysis of the factors and determine the fuzzy comprehensive evaluation method of mathematical statistics quantitative to qualitative indicators. In the modeling process, by changing the model number of neurons in the hidden layer and training function of class methods, respectively, compared to fit predicted ,from the predicted results shows that when using the BP neural network model which has 14 neurons and the Levenbeg-Marquardt for the training function to test sample, the obtained prediction error is very small, indicating that the network model has higher accuracy of prediction can be used to predict the quota data.Then, combining "Yunnan Quota Regulations ", using the trained model to forecast consumption with normal working conditions, and analyze the prediction results.Finally,using the AHP method to calculate amplitude difference coefficient of budget quota which is used to expand the construction quota to budget quota, and then find the supplementary quota.