Research On Funds Allocation For Maintenance Of Expressway Network Level In Shanxi

In recent years,China’s highway management departments are faced with a huge road network and need to determine when and where to take which maintenance measures under limited budget conditions to ensure the service level of the road network.Therefore,how to scientifically make appropriate allocation decisions for funds in maintenance management is of great significance.This paper analyzes the relevant information of the Shanxi Provincial Expressway Network Database,including basic information,historical performance data,maintenance history,traffic conditions,climate and environmental information and other data.For the missing and abnormal performance data,imputation was made and the abnormal data was eliminated by the Isolation Forest anomaly detection method.The characteristics related to the use performance of the pavement are selected,mainly including the basic information of the pavement,the structure type,the climate environment,the traffic load and the maintenance history,and the K-means clustering algorithm is used to classify the characteristics.The gradient boosting regression tree algorithm(GBRT)was used to establish a road performance prediction model suitable for the Shanxi expressway network for PCI and RQI.The model is based on the basic characteristics of the selected road surface and the performance index values of the previous three years.One year’s performance index value.The model is trained using the ten-fold cross-validation method,and the prediction results of the model show that the predicted value and the true value have a strong linear correlation.Comparing the GBRT model with the traditional linear regression model,Bayesian model and elastic network regression model,the GBRT model outperforms other models on the three indicators of MAE,MSE and R2.Two types of 0-1 integer programming models are established to optimize the allocation of pavement maintenance funds,which are aiming at the minimum funds and aiming at the best overall performance of the road network.Based on the gray wolf optimization algorithm,a discretization method of the gray wolf position search space is proposed,and nonlinear convergence factors and reverse learning theory are introduced to improve it.An improved binary gray wolf optimization algorithm(HBGWO)is proposed and used Based on the fund allocation model of this article,a case study was carried out on the expressway network of Taijiu Company.Finally,aiming at the three performance level constraints and the three funding level constraints,respectively,HBGWO and genetic algorithm(GA)are used to solve two models for comparison.The experimental results show that HBGWO can make a reasonable application of budget funds in solving the two models.In addition,HBGWO’s funding allocation model solution under different funding level constraints shows better robustness than GA.The research in this article provides a certain reference value for highway management departments to carry out maintenance fund allocation decision work in the future.