Intelligent Maintenance Scientific Decision-making System for Highway Pavement and Its Decision-making Method

Abstract

The present invention provides an intelligent maintenance scientific decision-making system for highway pavement and its decision-making method. The data collection and preprocessing module collects multi-source data on the public network and the highway management intranet, and uses specific algorithms to clean, integrate, and reduce the dimension. Based on the Django framework of Python, a road automatic detection system is built. The rough set theory is used to determine the index weights and calculate the PQI and MQI values, and a highway technical condition assessment report is generated. The pavement performance prediction module integrates multi-source data and uses a composite kernel function and a Gaussian process regression model to predict the pavement performance. The maintenance decision-making module determines the maintenance priority through ANP, generates a maintenance requirement list in combination with the hidden Markov model, and then uses the cuckoo search algorithm to generate a fund allocation plan. The maintenance benefit evaluation module uses the grey ideal solution method and the super-efficiency DEA model to evaluate the maintenance benefit. Finally, using the WebGL-based 3D visualization technology and the Echarts-based interactive query and analysis function, the road condition information display and auxiliary decision-making are realized.

Description

Technical Field

[0001] The present invention relates to the field of intelligent maintenance of highway pavements, and particularly to an intelligent maintenance scientific decision-making system for highway pavements and its decision-making method. Background Art

[0002] As an important part of the national transportation infrastructure, the pavement condition of expressways is directly related to the safety, comfort and efficiency of transportation. However, with the increase in the service life of expressways and the continuous growth of traffic flow, various diseases will appear on the pavement, such as cracks, potholes, ruts, etc., which seriously affect the service performance and quality of the pavement. The traditional decision-making methods for highway pavement maintenance mainly rely on manual experience and regular inspections, and there are problems such as low inspection efficiency, poor data accuracy, and lack of scientific decision-making.

[0003] On the one hand, manual inspection is limited by the professional level and work experience of inspectors, and it is difficult to comprehensively and accurately grasp the actual condition of the pavement. Moreover, the inspection cycle of manual inspection is long, and the development and change of pavement diseases cannot be detected in time, resulting in the lag of maintenance decision-making. On the other hand, traditional maintenance decision-making often only considers a single factor, such as the degree of pavement damage, while ignoring the comprehensive influence of various factors such as traffic volume, climate conditions, and maintenance costs, making the formulation of maintenance plans unreasonable and difficult to achieve the efficient use of maintenance funds and the maximization of pavement performance improvement.

[0004] In addition, with the rapid development of information technology, a large amount of pavement inspection data, traffic data, climate data, etc. are continuously accumulated, but there is a lack of effective data analysis and processing means, and these data have not fully played their role in maintenance decision-making. At the same time, the traditional maintenance information display method is relatively single, and it is difficult for decision-makers to intuitively understand the overall condition of the pavement and maintenance needs, which is not conducive to making scientific and reasonable decisions. Therefore, there is an urgent need for an intelligent and scientific intelligent maintenance scientific decision-making method for highway pavements to improve maintenance efficiency, reduce maintenance costs, and improve pavement service performance and quality. Summary of the Invention

[0005] The main purpose of the present invention is to provide an intelligent maintenance scientific decision-making system for highway pavements and its decision-making method, which solves the problem that the traditional maintenance information display method is relatively single, and it is difficult for decision-makers to intuitively understand the overall condition of the pavement and maintenance needs, which is not conducive to making scientific and reasonable decisions.

[0006] To solve the above technical problems, the technical solution adopted by the present invention is: an intelligent maintenance scientific decision-making method for highway pavements, the method includes: S1, collecting multi-source data of basic data, auxiliary basic data and maintenance decision data on the public network and the highway management intranet; preprocessing the above multi-source data;

[0007] S2. Build a road automatic detection system, input road damage, evenness, rut, and transverse force coefficient in the road sections to be maintained into the road automatic detection system, and the road automatic detection system generates a highway technical condition assessment report by combining PQI and MQI values;

[0008] S4. Based on the highway technical condition assessment report in step S2, combined with historical detection data, traffic volume data, and climate data, select a Gaussian process regression model to predict pavement performance and generate a pavement performance prediction curve for the next 1 - 5 years;

[0009] S4. Obtain the prediction curve, maintenance standard, and capital budget data from the pavement performance prediction curve in step S3, and use ANP to determine the maintenance priorities for different road sections;

[0010] Combined with the priorities determined by ANP, apply the hidden Markov model to predict the change trend of maintenance requirements and generate a maintenance requirements list;

[0011] S5. According to the maintenance requirements list in step S4, use the cuckoo search algorithm to generate a capital allocation plan, and obtain the maintenance plan, implementation cost, and expected effect data;

[0012] S6. Input the maintenance plan, implementation cost, and expected effect data into the analysis system, evaluate the improvement amplitude of pavement performance before and after maintenance, and the economic and social benefits of maintenance, and generate a maintenance benefit analysis report;

[0013] S7. Use the 3D visualization technology based on WebGL to display the road condition distribution, disease location, and maintenance plan information for the highway technical condition assessment report, pavement performance prediction curve, maintenance requirements list, and capital allocation plan in S2 - S5;

[0014] Then use the interactive query and statistical analysis function based on Echarts.

[0015] In the preferred solution, use the density - based spatial clustering algorithm to clean and integrate the multi - source data to remove noise data; use the locally linear embedding algorithm for data dimensionality reduction processing;

[0016] Among them, the basic data includes: map GIS data, road network information data, highway road surface and subgrade data, and pavement technical condition assessment data;

[0017] The auxiliary basic data includes: traffic flow data, meteorological data, geological environment data, and administrative management data;

[0018] Maintenance decision - making data: maintenance material category data, maintenance equipment data, maintenance method data, and maintenance project unit price data.

[0019] In the preferred solution, a road automation detection system is built using the Django framework based on Python;

[0020] Use the django - admin startproject command to create a Django project, and then execute the python manage.py startapp command to create a specific application;

[0021] Configure the database connection information in the project's configuration file, and select a MySQL or SQLite database; define the database models for road detection data and index calculation. The database models include road damage, flatness, rutting, transverse force coefficient detection data, and PQI and MQI index data. Execute the python manage.py makemigrations and python manage.py migrate commands to create the corresponding database tables for the database models;

[0022] Write a Python script to collect road damage, flatness, rutting, and transverse force coefficient data for the section to be maintained; use Django's REST framework to develop API interfaces to receive the collected data and store it in the database;

[0023] Write Python code in the Django application to implement the rough set theory algorithm, and construct an information system with the collected data , where is the set of objects, , where is the set of conditional attributes, i.e., each detection index, is the set of decision attributes, i.e., the overall performance of the highway, is the set of attribute values, is the information function;

[0024] Perform attribute reduction by calculating the discernibility matrix, and the formula is:

[0025] ( );

[0026] Calculate the attribute importance , where , ;

[0027] Finally, determine the weights of each index according to ;

[0028] Standardize the original index data of PCI, RQI, RDI, and PBI that make up PQI collected again;

[0029] Among them, PCI is the pavement damage condition index, RQI is the pavement riding quality index, RDI is the pavement rut depth index, and PBI is the pavement skid resistance performance index;

[0030] Then, according to calculate the PQI value;

[0031] For MQI, perform the same standardization processing on the original data of its constituent indicators PQI, BCI, TCI, and SCI. Again, use the rough set theory to determine the weights of each indicator, and then according to calculate the MQI value;

[0032] Among them, PQI is the pavement service performance index, and MQI is the highway technical condition index;

[0033] According to the calculated PQI and MQI values, combined with the pre-set evaluation grade standards, generate a highway technical condition evaluation report including road section basic information, index calculation results, evaluation grades, and analysis and maintenance suggestions in the Django application;

[0034] Use the Matplotlib and Echarts libraries to visually display the evaluation results; develop a Web interface using the Django template system to view data and evaluation report information;

[0035] Use the Django test framework to test each functional module of the system to ensure the normal operation of data collection, calculation, and report generation functions;

[0036] Deploy the system to a cloud server, and use Nginx as the Web server and Gunicorn as the application server.

[0037] In the preferred solution, the specific steps of step S3 are as follows:

[0038] S31. Collect the PQI, MQI, and each sub-item index values in the highway technical condition evaluation report in step S2, and integrate them with the collected historical detection data, traffic volume data, and climate data; use the method based on local outlier factor to detect outliers. For the sample points in the dataset, calculate the local reachability density through the formula calculate the local outlier factor through the formula When is greater than the set threshold, process the outliers;

[0039] Among them is the set of nearest neighbor samples of , is the sample to the distance to its nearest neighbor, which is the Euclidean distance between the sample and ;

[0040] S32. Adopt a composite kernel function , for the training data set , calculate the covariance matrix , whose elements , considering the observation noise to obtain the covariance matrix with noise ;

[0041] S33. Define the log-likelihood function of Gaussian process regression , and use the Bayesian optimization algorithm to find the optimal hyperparameters ;

[0042] S34. Generate prediction input data according to the future time span of 1 - 5 years and the historical data change trend , calculate the covariance matrix between the prediction input and the training data and the covariance matrix of the prediction input itself , calculate the prediction mean through the formula , and calculate the prediction covariance through the formula ;

[0043] S35. Calculate the confidence interval of the predicted value based on the prediction covariance matrix , and plot the changes of the prediction mean and the confidence interval over time to form a pavement performance prediction curve within the next 1 - 5 years.

[0044] In the preferred solution, the specific steps of step S4 are as follows:

[0045] S41. Collect the pavement performance prediction curve data, maintenance standards, and capital budget data of step S3; use the method based on the interquartile range to process outliers, and determine the outlier judgment interval through the formula , and correct the outliers using linear interpolation; use the min - max normalization formula to normalize the pavement performance prediction curve data;

[0046] S42. Construct a network structure model including element sets of pavement performance, traffic impact, and economic cost, and analyze the interdependent relationship between elements; construct a judgment matrix through the expert scoring method ; Solve to obtain the eigenvector, and combine it into an unweighted supermatrix ; Determine the element set correlation weight matrix ​, calculate the weighted supermatrix ; for perform power operation to obtain the limit supermatrix , determine the maintenance priority;

[0047] S43. Define the set of maintenance demand states and the observation set ; Initialize the state transition probability matrix , the observation probability matrix and the initial state probability vector ; Use the Baum - Welch algorithm, through the forward probability recurrence formula and the backward probability recurrence formula to train the model and update the parameters ; Use the Viterbi algorithm, through the recurrence formula to find the most likely state sequence;

[0048] S44. Combine the priority determined by ANP and the trend predicted by the hidden Markov model, generate a maintenance demand list including road section information, maintenance time, maintenance demand status, and estimated required funds in the order of road section number and time, and adjust and screen maintenance projects according to the funding budget formula

[0049] In the optimal solution, the specific steps of step S5 are as follows:

[0050] S51. Extract the road section maintenance demand information, estimated required funds and the maintenance priority from the maintenance demand list in step S4, clarify the total funding budget and the annual funding allocation ratio ; Define the objective function , where:

[0051] Define the maintenance benefit function as ;

[0052] Define the maintenance cost function as: ;

[0053] Set the constraint conditions and the priority penalty term ;

[0054] S52. In the steps of the cuckoo search algorithm, select the solution with the largest objective function value as the optimal solution;

[0055] Generate a funding allocation plan according to the optimal solution, determine the maintenance road sections and the allocated funds, clarify the maintenance plan; and calculate the total implementation cost and the expected effect data 。

[0056] In the preferred solution, in step S6, the maintenance plan, implementation cost, and expected effect data are input into the analysis system, and the analysis system uses the grey ideal solution method to evaluate the improvement amplitude of the pavement performance before and after maintenance;

[0057] The analysis system then uses the super-efficiency DEA model to evaluate the economic and social benefits of the maintenance, and generates a maintenance benefit analysis report.

[0058] In the preferred solution, the specific steps of step S6 are as follows:

[0059] S61. Collect the maintenance plan, implementation cost, and expected effect data in steps S51 and S52 and input them into the analysis system; perform standardization processing on the original data matrix to eliminate the influence of dimensions;

[0060] S62. Determine the standardized data matrix and calculate the pavement performance improvement amplitude index;

[0061] S63. Use the super-efficiency DEA model to evaluate the economic and social benefits:

[0062] Take each maintenance plan as a decision-making unit, determine the input implementation cost and the output economic and social benefit indicators; establish a super-efficiency DEA model, and the objective function is ;

[0063] The constraint conditions are:

[0064] , , Solve to obtain the super-efficiency value ;

[0065] where is the th input index value of the th DMU, is the th output index value of the th DMU, and are the weights of the output and input indicators respectively, is a non-Archimedean infinitesimal;

[0066] S64. The report includes a cover and table of contents, project overview, pavement performance improvement evaluation 、economic and social benefit evaluation 、comprehensive evaluation and suggestions, and appendix.

[0067] In the optimal solution, integrate the highway technical condition assessment report, pavement performance prediction curve, maintenance requirement list, and fund allocation plan in steps S2 - S5; convert the road section geographical location information into geographical coordinates suitable for WebGL rendering;

[0068] Use the WebGL library to create a 3D scene and load the geographical base map; create a 3D model of the road section according to the road section geographical location, and use the formula Mark the disease location with a specific geometric shape;

[0069] where is the base size, is the standardized disease degree data, is the scaling factor, used to control the variation range of the disease mark size;

[0070] According to the maintenance requirement list and fund allocation plan, represent the maintenance measures with different - colored lines or textures, and use the formula Show the maintenance progress in animation;

[0071] where and are the start color and end color respectively, progress is the completion progress of the maintenance plan;

[0072] Bind the integrated data with Echarts, create bar charts, line charts, and pie charts; add interactive events and query input boxes to the charts to achieve interactive query; provide statistical analysis functions such as summation and average value, and display the statistical results with new charts or tables; set the data update period to achieve dynamic data update;

[0073] Use the formula to calculate the comprehensive benefit index;

[0074] where 、 and are the weight coefficients, determined by expert scoring or historical data statistics, is the improvement value of the pavement performance PQI, cost is the maintenance cost, is the quantified value of social benefits;

[0075] Provide decision - making suggestions for adjusting the maintenance plan and optimizing the fund allocation for decision - makers according to the PQI and MQI indicators.

[0076] In the optimal solution, it includes a data collection and pre - processing module, a road automatic detection system, a pavement performance prediction module, a maintenance decision - making module, a maintenance benefit evaluation module, and a visualization and analysis module;

[0077] The data collection and preprocessing module collects multi-source data of basic data, auxiliary basic data and maintenance decision data from the public network and highway management intranet, and uses density-based spatial clustering algorithm for cleaning and integration, and adopts local linear embedding algorithm for data dimension reduction processing;

[0078] The road automation detection system is built with the Python-based Django framework, which uses the rough set theory algorithm to determine the detection indicator weights and calculate the PQI and MQI values for data collection, indicator calculation, report generation and display, system testing and deployment functions;

[0079] The pavement performance prediction module is used to integrate relevant data, process outliers and construct derived features. It uses a composite kernel function and Gaussian process regression model combined with a Bayesian optimization algorithm to predict pavement performance and generate a pavement performance prediction curve for the next 1 to 5 years.

[0080] The maintenance decision module is used to determine maintenance priorities, predict the trend of maintenance demand changes to generate a maintenance demand list, use the cuckoo search algorithm to generate a funding allocation plan, and determine the maintenance plan, implementation cost, and expected effect data;

[0081] The maintenance benefit evaluation module uses the grey ideal solution and the super-efficiency DEA model to evaluate the improvement of pavement performance before and after maintenance, economic benefits and social benefits, and generates a maintenance benefit analysis report;

[0082] The visualization and analysis module uses WebGL-based 3D visualization technology to display road condition distribution, disease location, and maintenance plan information. It uses Echarts-based interactive query and statistical analysis functions to support dynamic data updates and calculate comprehensive benefit indicators to assist decision-making.

[0083] The present invention provides a scientific decision-making system for intelligent maintenance of highway pavement and its decision-making method. At the data processing and analysis level, it can efficiently integrate multi-source data such as highway technical condition assessment reports, pavement performance prediction curves, maintenance needs lists and funding allocation plans, and process them through advanced algorithms to improve data accuracy and availability and tap the potential value of data.

[0084] In terms of maintenance decision-making, rough set theory is used to determine indicator weights, combined with the Gaussian process regression model to predict road performance, maintenance priorities are determined based on ANP, and then the cuckoo search algorithm is used to optimize fund allocation, so that decisions take into account multiple factors and are more scientific and reasonable, thereby improving the efficiency of the use of maintenance funds and accurately investing resources in the sections that need the most maintenance.

[0085] In terms of evaluation and report generation, the grey ideal solution method and the super-efficiency DEA model are used to evaluate the maintenance benefits, and a comprehensive maintenance benefit analysis report is generated to provide strong data support and clear evaluation results for decision-making.

[0086] In the field of visualization and interactive analysis, the 3D visualization technology based on WebGL intuitively displays the road condition distribution, disease locations, and maintenance plans. Combining with the interactive query and statistical analysis functions of Echarts, it supports dynamic data updates, facilitating decision-makers to intuitively grasp the road conditions, obtain detailed information in a timely manner, and conduct statistical analysis, thus making scientific decisions quickly.

[0087] This method can improve the intelligent and scientific level of highway pavement maintenance, reduce maintenance costs, improve pavement performance and service quality, ensure road traffic safety and smoothness, and promote the development of highway maintenance management towards high efficiency and precision. Brief Description of the Drawings

[0088] The present invention will be further described below in conjunction with the drawings and embodiments:

[0089] Figure 1 It is a diagram of the intelligent maintenance scientific decision-making system for highway pavement of the present invention;

[0090] Figure 2 It is a diagram of the data collection and preprocessing module of the present invention;

[0091] Figure 3 It is a diagram of the road automatic detection system of the present invention;

[0092] Figure 4 It is a diagram of the pavement performance prediction module of the present invention;

[0093] Figure 5 It is a diagram of the maintenance decision-making module of the present invention;

[0094] Figure 6 It is a diagram of the maintenance benefit evaluation module of the present invention;

[0095] Figure 7 It is a diagram of the visualization and analysis module of the present invention. Detailed Embodiments

[0096] Embodiment 1

[0097] As Figures 1-7 shown, a method for intelligent maintenance scientific decision-making of highway pavement, the method includes:

[0098] S1. Collect multi-source data of basic data, auxiliary basic data, and maintenance decision data on the public network and the highway management intranet; preprocess the above multi-source data;

[0099] S2. Set up a road automatic detection system, input road damage, evenness, rut, and transverse force coefficient in the sections to be maintained into the road automatic detection system, and the road automatic detection system generates a highway technical condition assessment report by combining PQI and MQI values;

[0100] S3. Based on the highway technical condition assessment report in step S2, combine historical detection data, traffic volume data, and climate data, select the Gaussian process regression model to predict pavement performance, and generate a pavement performance prediction curve for the next 1 - 5 years;

[0101] S4. Obtain the prediction curve, maintenance standard, and capital budget data according to the pavement performance prediction curve in step S3, and use ANP to determine the maintenance priority of different sections for the prediction curve, maintenance standard, and capital budget data;

[0102] Combined with the priority determined by ANP, apply the hidden Markov model to predict the change trend of maintenance demand and generate a maintenance demand list;

[0103] S5. According to the maintenance demand list in step S4, use the cuckoo search algorithm to generate a capital allocation plan, and obtain the maintenance plan, implementation cost, and expected effect data;

[0104] S6. Input data such as the maintenance plan, implementation cost, and expected effect into the analysis system, evaluate the improvement range of pavement performance before and after maintenance, the economic and social benefits of maintenance, and generate a maintenance benefit analysis report;

[0105] S7. Use the 3D visualization technology based on WebGL to display the road condition distribution, disease location, and maintenance plan information for the highway technical condition assessment report, pavement performance prediction curve, maintenance demand list, and capital allocation plan in S2 - S5;

[0106] Then use the interactive query and statistical analysis function based on Echarts to support dynamic data update and assist in decision - making.

[0107] In the preferred solution, use the density - based spatial clustering algorithm to clean and integrate the multi - source data and remove noise data; use the locally linear embedding algorithm for data dimensionality reduction processing;

[0108] Among them, the basic data includes: map GIS data, road network information data, highway pavement and subgrade data, and pavement technical condition assessment data;

[0109] The auxiliary basic data includes: traffic flow data, meteorological data, geological environment data, and administrative management data;

[0110] Maintenance decision - making data: maintenance material category data, maintenance equipment data, maintenance method data, and maintenance project unit price data.

[0111] Comprehensively collect necessary basic data such as map GIS data, road network information data, highway pavement and subgrade data, pavement technical condition evaluation data, etc., auxiliary basic data such as traffic flow data, meteorological data, geological environment data, administrative management data, etc., and maintenance decision-related data such as maintenance material category data, maintenance equipment data, maintenance method data, maintenance project unit price data, etc. Use the density-based spatial clustering algorithm (DBSCAN) to clean and integrate multi-source heterogeneous data, identify and remove noise data such as outliers, and achieve effective integration of data. Use the locally linear embedding algorithm (LLE) for data dimensionality reduction, reduce the redundant data dimension while preserving the local geometric structure of the data, reduce the data complexity, improve the subsequent calculation efficiency, and provide high-quality data support for subsequent decisions.

[0112] Example 2

[0113] Further illustrate in combination with Example 1, and build a road automatic detection system using the Django framework based on Python;

[0114] Use the django - admin startproject command to create a Django project, and then execute the python manage.py startapp command to create a specific application;

[0115] Configure the database connection information in the project's configuration file, and you can choose a MySQL or SQLite database; define the database models related to road detection data and index calculation. The database models include road damage, flatness, rut, and transverse force coefficient detection data, as well as PQI and MQI index data. Execute the python manage.py makemigrations and python manage.py migrate commands to create the corresponding database tables for the database models;

[0116] Write a Python script to collect road damage, flatness, rut, and transverse force coefficient data of the road section to be maintained; use the REST framework of Django to develop API interfaces, receive the collected data and store it in the database;

[0117] Write Python code in the Django application to implement the rough set theory algorithm, and construct an information system with the collected data , where is the object set, , where is the conditional attribute set, that is, each detection index, is the decision attribute set, that is, the overall performance of the highway, is a set of attribute values, is an information function;

[0118] Attribute reduction is performed by calculating the discernibility matrix, and the formula is ( );

[0119] Calculate the importance of attributes , where , ;

[0120] Finally, according to Determine the weights of each index;

[0121] Standardize the original data of the indexes that make up PQI, namely PCI, RQI, RDI, and PBI, collected;

[0122] Among them, PCI is the pavement damage condition index, RQI is the pavement riding quality index, RDI is the pavement rut depth index, and PBI is the pavement skid resistance performance index;

[0123] Then, according to Calculate the PQI value; for MQI, standardize the original data of its constituent indexes PQI, BCI, TCI, and SCI in the same way, determine the weights of each index by applying the rough set theory again, and then according to Calculate the MQI value;

[0124] According to the calculated PQI and MQI values, combined with the pre-set evaluation grade standard, generate a highway technical condition evaluation report including road section basic information, index calculation results, evaluation grade, and analysis and maintenance suggestions in the Django application;

[0125] Use the Matplotlib and Echarts libraries to visually display the evaluation results; develop a Web interface using the Django template system to view data and evaluation report information;

[0126] Use the Django test framework to test each functional module of the system to ensure that the data collection, calculation, and report generation functions are normal; deploy the system to a cloud server, and use Nginx as the Web server and Gunicorn as the application server.

[0127] In the above: The acquisition methods of road damage, evenness, rut, and lateral force coefficient are as follows:

[0128] The road damage coefficient is obtained by using advanced equipment installed on the inspection vehicle, such as high-definition cameras and laser scanners, to quickly collect pavement images and three-dimensional data during vehicle driving. Through image processing and analysis techniques, identify and classify road damage and calculate its relevant parameters.

[0129] The flatness coefficient is measured by using a laser sensor to measure the change in the distance between the road surface and the sensor, and the road surface flatness index is obtained by processing the data with a computer. This method can measure the flatness of the road surface quickly and accurately, and can also obtain the three-dimensional information of the road surface simultaneously.

[0130] The rut coefficient is measured by a laser rut meter installed on the inspection vehicle. During the vehicle's driving, the cross-section of the road surface is scanned by laser to quickly and accurately measure parameters such as the depth, width, and shape of the rut, and the data is automatically recorded and processed.

[0131] The transverse force coefficient is measured by a special friction coefficient measuring vehicle. When driving at a certain speed, the transverse force coefficient is measured by using the frictional force between the test wheel and the road surface. The test wheel usually uses a special tire and maintains a certain vertical load and slip ratio during the test.

[0132] Example 3

[0133] Further illustrated in combination with Examples 1-2, the specific steps of step S3 are as follows:

[0134] S31. Collect the PQI, MQI, and the values of each sub-item index in the highway technical condition assessment report in step S2, and integrate them with the collected historical detection data, traffic volume data, and climate data; use the method based on the local outlier factor to detect outliers. For the sample points in the dataset , calculate the local reachability density through the formula , calculate the local outlier factor through the formula . When is greater than the set threshold, process the outliers; construct derivative features, where: the moving average of the traffic volume , the comprehensive index of climate factors , where is determined by principal component analysis;

[0135] S32. Use the composite kernel function . For the training dataset , calculate the covariance matrix , whose elements . Considering the observation noise to obtain the covariance matrix with noise ;

[0136] S33. Define the log-likelihood function of Gaussian process regression , and use the Bayesian optimization algorithm to find the optimal hyperparameters ;

[0137] S34. Generate prediction input data according to the time span of the next 1-5 years and the changing trend of historical data , calculate the covariance matrix between the predicted input and the training data and the covariance matrix of the predicted input itself , through the formula calculate the predicted mean, and through the formula calculate the predicted covariance;

[0138] S35. Based on the predicted covariance matrix calculate the confidence interval of the predicted value, and plot the changes of the predicted mean and the confidence interval over time to form a pavement performance prediction curve for the next 1 - 5 years.

[0139] The specific steps are as follows:

[0140] Collect the PQI, MQI and the values of each sub - index in the highway technical condition assessment report in step S2, historical detection data (pavement performance - related indicators over the years), traffic volume data (traffic flow and vehicle type ratio at different times), and climate data (temperature, precipitation, sunshine duration, etc.). Integrate these data into a multi - dimensional dataset according to the time series and road section information.

[0141] Adopt the method based on the Local Outlier Factor (LOF) to detect outliers. For each sample point in the dataset , calculate its local reachability density and the local outlier factor .

[0142] Local reachability density formula:

[0143] where is 's set of k - nearest neighbor samples, , is the distance from sample to its -th nearest neighbor, is the Euclidean distance between sample and .

[0144] Local outlier factor formula:

[0145] When is greater than the set threshold (such as 1.5), it is considered that is an outlier, and interpolation or deletion methods can be used for processing. The purpose of this method is to remove the noise points in the data and avoid their adverse effects on model training.

[0146] In addition to the original data features, construct some derivative features. Calculate the moving average of the traffic volume , where is the traffic volume at the moment, is the window size of the moving average. A comprehensive index of climate factors can also be constructed, such as , where are the weight coefficients determined by principal component analysis (PCA). These derived features help the model capture the potential patterns in the data.

[0147] A composite kernel function is adopted, which combines the radial basis function (RBF) and the rational quadratic kernel function (RQ). The formula for the composite kernel function is: ;

[0148] where and are the signal variances of the RBF and RQ kernel functions respectively, and are their length scales respectively, is the shape parameter of the rational quadratic kernel function. The composite kernel function can better adapt to different types of data features and improve the flexibility of the model.

[0149] For the training dataset , the elements of the covariance matrix are . At the same time, considering the observation noise , the covariance matrix with noise is obtained, where is the identity matrix. The covariance matrix describes the correlation between data points and is the core of Gaussian process regression.

[0150] The log-likelihood function of Gaussian process regression is: ;

[0151] where is the vector of observed values of the training data, is the vector of hyperparameters of the model, is the determinant of. This likelihood function measures the probability of the observed data given the hyperparameters.

[0152] The Bayesian optimization algorithm is used to find the optimal hyperparameters . Bayesian optimization estimates the posterior distribution of the objective function (here the log-likelihood function) by constructing a surrogate model (such as a Gaussian process model), and then selects the next hyperparameter point to be evaluated according to the acquisition function (such as expected improvement EI). This process is repeated until the optimal hyperparameter combination is found. Bayesian optimization can find better hyperparameters with fewer evaluation times and improve the optimization efficiency.

[0153] Generate prediction input data based on the time span of the next 1 - 5 years, combined with the historical traffic volume and the changing trend of climate data For the traffic volume, a time series prediction method can be used for prediction, and the predicted value is used as the input

[0154] Calculate the covariance matrix between the prediction input and the training data , element , and the covariance matrix of the prediction input itself .

[0155] Prediction mean formula:

[0156] Prediction covariance formula:

[0157] The prediction mean gives the predicted value of the pavement performance, and the prediction covariance reflects the uncertainty of the prediction

[0158] Based on the prediction covariance matrix , calculate the confidence interval of the predicted value. The 95% confidence interval is , where is the diagonal element of. This helps to evaluate the reliability of the prediction result

[0159] Plot the changes of the prediction mean and the confidence interval over time to form a pavement performance prediction curve within the next 1 - 5 years. This curve visually shows the changing trend and the uncertainty range of the pavement performance, providing a scientific basis for highway maintenance decisions

[0160] Example 4

[0161] Further illustrate in combination with Examples 1 - 3: The specific steps of step S4 are as follows

[0162] S41. Collect the pavement performance prediction curve data, maintenance standards, and capital budget data of step S3; use the method based on the interquartile range to process outliers, and determine the outlier judgment interval through the formula , and correct the outliers using linear interpolation; use the min - max normalization formula to normalize the pavement performance prediction curve data ,

[0163] S42. Construct a network structure model including element sets of pavement performance, traffic impact, and economic cost, and analyze the interdependent relationship between elements; construct a judgment matrix through the expert scoring method ; Solve to obtain the eigenvector, and combine it into an unweighted supermatrix ; Determine the associated weight matrix of the element set , calculate the weighted supermatrix ; perform power operation on to obtain the limit supermatrix , and determine the maintenance priority;

[0164] S43. Define the set of maintenance demand states and the set of observations ; initialize the state transition probability matrix , the observation probability matrix and the initial state probability vector ; use the Baum - Welch algorithm, through the forward probability recurrence formula and the backward probability recurrence formula to train the model and update the parameters ; use the Viterbi algorithm, through the recurrence formula to find the most likely state sequence;

[0165] S44. Combine the priority determined by ANP and the trend predicted by the hidden Markov model, generate a maintenance demand list including road section information, maintenance time, maintenance demand status, and estimated required funds in the order of road section number and time, and adjust and screen maintenance projects according to the funding budget formula

[0166]

[0167] The specific implementation method is as follows:

[0167] Extract the road surface performance prediction curve data for the next 1 - 5 years of each road section from step S3, covering key indicators such as PQI and MQI. At the same time, collect detailed maintenance standards, maintenance measures and quality requirements corresponding to different PQI and MQI intervals; obtain funding budget data to clarify the total funds available for maintenance and the annual funding allocation ratio ( ).

[0168] Check the road surface performance prediction curve data, identify and process outliers. For outliers, use the method based on the interquartile range (IQR). Let the lower quartile of the data be , the upper quartile be , then IQR = . The outlier judgment interval is , and the values outside this interval are regarded as outliers, and linear interpolation can be used for correction. Normalize the road surface performance prediction curve data.

[0169] Divide the factors affecting the maintenance priority into multiple element sets. Include the road surface performance element set and the traffic impact element set Economic cost element set .

[0170] Analyze the interdependencies within and between each element set. For example, pavement performance affects traffic flow, and traffic flow in turn affects maintenance costs. Construct a network structure to visually display these relationships.

[0171] For element pairs with associations in the network, construct a judgment matrix through the expert scoring method. For the elements in the element set and the element , the judgment matrix , where represents the degree of importance of element relative to element . Determine its value using the 1 - 9 scale method. For each judgment matrix , calculate its eigenvector , obtained by solving , where is the largest eigenvalue of matrix . Combine the eigenvectors corresponding to each judgment matrix into an unweighted supermatrix . The unweighted supermatrix reflects the relative importance between elements but does not consider the association weights between element sets. Determine the association weight matrix between element sets, also obtained through expert scoring and eigenvector calculation. The weighted supermatrix . The weighted supermatrix comprehensively considers the relationship weights within and between element sets.

[0172] Perform a power operation on the weighted supermatrix . When is large enough, converges to the limit supermatrix . The sum of the elements in each column of the limit supermatrix is 1, and its element values represent the relative importance of each road section under the influence of all factors considered, that is, the maintenance priority. The power operation formula is .

[0173] According to the maintenance standards, define the maintenance demand status set , representing "no maintenance required", "light maintenance", "medium maintenance", and "heavy maintenance" respectively.

[0174] Select observable variables closely related to maintenance needs as the observation set , such as the normalized PQI, traffic flow, and temperature (considering the impact of climate on pavement performance).

[0175] State transition probability matrix :[[]] , which represents the probability of transitioning to state at time when in state . The initial value can be estimated based on the statistical results of historical maintenance data. The observation probability matrix : , which represents the probability of observing at time when in state . It is also initialized according to historical data. The initial state probability vector

[0176] : , which represents the probability of being in state at the initial time, and is determined based on the initial pavement performance prediction data. The forward probability and the backward probability are introduced, where are model parameters. The forward probability recurrence formula:

[0177] , which is used to calculate the probability of being in state at time and observing the first observations. The backward probability recurrence formula: , which is used to calculate the probability of being in state at time and observing the subsequent observations. By continuously iterating and updating , and , the likelihood probability of the observation sequence is maximized to optimize the model parameters.

[0178] Using the trained hidden Markov model and observation data such as pavement performance prediction curves, the most likely state sequence is found through the Viterbi algorithm. The Viterbi algorithm is based on the idea of dynamic programming. Define to represent the maximum probability of being in state at time and observing the first observations. The recurrence formula is , and the backtracking pointer is recorded. Finally, the most likely state sequence is backtracked to obtain the changing trend of maintenance requirements.

[0179] Combining the maintenance priorities determined by ANP and the changing trend of maintenance requirements predicted by the hidden Markov model, a maintenance requirement list is generated according to the section numbers and time order. The list content includes:

[0180] Section numbers, section lengths, starting and ending positions, etc.

[0181] Determine the specific time points for maintenance of each section according to the predicted trend of maintenance demand changes.

[0182] Clarify the maintenance demand status of each section at the corresponding time point, including light maintenance and medium maintenance.

[0183] Estimate the funds required for each maintenance of each section according to the maintenance standards and maintenance demand status . At the same time, combined with the fund budget data, check whether the total maintenance funds for each year exceed the fund allocation quota for that year . If it exceeds, adjust and screen the maintenance projects according to the maintenance priority.

[0184] Embodiment 5

[0185] Further illustrate in combination with Embodiments 1-4. The specific steps of Step S5 are as follows:

[0186] S51. Extract the section maintenance demand information, estimated required funds and maintenance priority from the maintenance demand list in Step S4, clarify the total fund budget and the annual fund allocation ratio ; Define the objective function , where ( ), ; Set the constraint conditions and the priority penalty term ;

[0187] S52. Steps of the cuckoo search algorithm: Randomly generate binary vectors of length as the initial solution; Set the discovery probability , the maximum number of iterations , the step size scaling factor parameters; Use the Lévy flight formula to generate the step size , and generate a new solution through ; Calculate the objective function value of and check the constraint conditions. When not satisfied, correct the solution; Compare and . If then use to replace ; Update the nest randomly with probability to generate a new solution And evaluate the replacement;

[0188] Terminate the algorithm when the maximum number of iterations is reached ; Select the solution with the largest objective function value as the optimal solution ; According to Generate a fund allocation plan, determine the maintenance sections and allocated funds; clarify the maintenance plan; calculate the implementation cost and expected effect data .

[0189] The specific implementation method is as follows:

[0190] Extract the maintenance requirement information of each section from the maintenance requirement list in step S4, including the required maintenance type (including light maintenance, medium maintenance, heavy maintenance), the estimated required funds ( represents the section number) and the maintenance priority determined by ANP . At the same time, clarify the total budget of funds and the fund allocation ratio for each year ( ).

[0191] Define the maintenance benefit function , where is the expected benefit improvement value after the section is maintained, which can be calculated according to the improvement range of indicators such as PQI and MQI in the pavement performance prediction curve. For example ( and are weight coefficients, determined by expert scoring or historical data statistics), is a decision variable, , represents that the section is maintained, represents that it is not maintained.

[0192] Define the maintenance cost function . Considering comprehensively, the objective function can be expressed as , where and are weight coefficients for weighing the maintenance benefit and cost, which can be adjusted according to the actual situation. , ensuring that the allocated funds do not exceed the total budget.

[0193] For sections with higher maintenance priorities, give priority to ensuring that they have a higher value possibility. A priority penalty term can be introduced and minimized as much as possible during the optimization process.

[0194] Randomly generate initial solutions (cuckoo nests), where each solution is a binary vector of length ( ), where represents whether section in the th solution is to be maintained.

[0195] Set the parameters of the algorithm, including the discovery probability (used to control the probability that a cuckoo nest is discovered and replaced), the maximum number of iterations , and the step size scaling factor . For each cuckoo nest , generate a new solution using Lévy flight. The step size of Lévy flight is calculated by the formula , where and are random variables following a normal distribution, and is usually taken as 1.5. The formula for generating a new solution is , where is a random vector whose elements take values and are used to randomly change some decision variables in the solution.

[0196] Calculate the objective function value of the new solution , and at the same time check whether it satisfies the constraint conditions. If the total funds limit or the maintenance priority constraint is not satisfied, the solution is corrected, for example, randomly reducing some sections until the constraint conditions are met.

[0197] Compare the objective function values of the new solution and the original solution . If , then replace with . Update some nests randomly with probability . For the selected nests , randomly generate a new solution , and also evaluate its objective function value and the situation of satisfying the constraint conditions. If it is better than the original solution, replace it.

[0198] When the maximum number of iterations is reached, the algorithm terminates.

[0199] Select the solution with the largest objective function value from all the cuckoo nests as the optimal solution .​

[0200] According to the optimal solution , determine which road sections need to be maintained and the funds allocated to each road section. For the road sections , allocate the estimated required funds .

[0201] Clarify the type of maintenance adopted for each road section to be maintained, which is determined according to the maintenance requirement list.

[0202] Calculate the total cost of implementing this maintenance plan .

[0203] According to the benefit calculation method in the objective function, calculate the expected benefit improvement value after implementing this maintenance plan .

[0204] Example 6

[0205] Combined with Examples 1-5 for further illustration, the specific steps of step S6 are as follows: input data such as the maintenance plan, implementation cost, and expected effect into the analysis system, and the analysis system uses the grey ideal solution method to evaluate the improvement amplitude of the pavement performance before and after maintenance;

[0206] The analysis system then uses the super-efficiency DEA model to evaluate the economic and social benefits of the maintenance, and generates a maintenance benefit analysis report.

[0207] In the optimal solution, S61. Collect data on the maintenance plan, implementation cost, and expected effect and input them into the analysis system; perform standardization processing on the original data matrix , for positive indicators, use , and for negative indicators, use formula to eliminate the influence of dimension;

[0208] S62. Determine the ideal solution of the standardized data matrix ( ) and the negative ideal solution ( ); calculate the grey correlation coefficient ( ) with the ideal solution and the grey correlation coefficient ( ) with the negative ideal solution; calculate the grey correlation degree and ; calculate the pavement performance improvement amplitude index ;

[0209] S63. Evaluate economic and social benefits using the super-efficiency DEA model: Take each maintenance plan as a decision-making unit, determine the input implementation cost and the output economic and social benefit indicators; establish a super-efficiency DEA model, with the objective function being , and the constraint conditions being , , , and solve to obtain the super-efficiency value ;

[0210] S64. The report includes a cover and table of contents, project overview, pavement performance improvement evaluation , economic and social benefit evaluation , comprehensive evaluation and suggestions, and appendix;

[0211] The specific implementation method is as follows:

[0212] Collect the detailed data of the maintenance plans, implementation costs, and expected effects obtained in step S5. The expected effects cover various pavement performance indicators before and after maintenance, such as pavement damage rate, smoothness, anti-skid performance, etc.; the implementation costs include expenses for labor, materials, equipment leasing, etc.; the social benefit data can include the degree of traffic congestion alleviation, satisfaction of surrounding residents, etc. Enter these data into the analysis system.

[0213] Since different indicators have different dimensions and value ranges, in order to make the indicators comparable, linear transformation is used for standardization. Let the original data matrix be , where is the number of maintenance plans, and is the number of indicators.

[0214] For positive indicators (the larger the value, the better, such as the improvement rate of pavement smoothness), the standardization formula is:

[0215] ;

[0216] For negative indicators (the smaller the value, the better, such as the implementation cost), the standardization formula is:

[0217] ;

[0218] Eliminate the influence of different indicator dimensions, enable the indicators to be compared on a unified scale, and provide a basis for subsequent evaluation.

[0219] Let the standardized data matrix be . The ideal solution is the vector composed of the optimal values of each indicator, that is, , where ; The negative ideal solution is the vector composed of the worst values of each indicator, that is, , where . It serves as a reference standard for evaluating the improvement of pavement performance under each maintenance plan.

[0220] Calculate the grey correlation coefficients of each maintenance plan with the ideal solution and the negative ideal solution. For the th maintenance plan and the th index, the grey correlation coefficient with the ideal solution is calculated as follows:

[0221] ;

[0222] where , is the discrimination coefficient, usually taken as 0.5.

[0223] The grey correlation coefficient with the negative ideal solution is calculated as follows:

[0224] ;

[0225] where . The above algorithm is used to measure the degree of closeness between the index values of each maintenance plan and the ideal solution and the negative ideal solution.

[0226] Perform a weighted sum of the grey correlation coefficients for each maintenance plan to obtain the grey correlation degree with the ideal solution and the grey correlation degree with the negative ideal solution. Let the index weight vector be , and , then:

[0227] ;

[0228] ;

[0229] Taking into account the influence of each index, obtain the overall correlation degree of each maintenance plan with the ideal solution and the negative ideal solution.

[0230] The pavement performance improvement amplitude index is calculated as follows:

[0231]

[0232] Function: Quantify the improvement amplitude of pavement performance before and after maintenance for each maintenance plan, the closer it is to 1, the greater the improvement amplitude of pavement performance.

[0233] Each maintenance plan is regarded as a decision-making unit (DMU). The input indicators are selected as the various costs related to the implementation cost, and the output indicators are selected as the economic benefits (such as the economic benefits brought by the improvement of the road traffic capacity after maintenance) and social benefits (such as the quantified value of traffic congestion mitigation, the improved value of residents' satisfaction, etc.).

[0234] For the th DMU, the objective function of its super-efficiency DEA model is:

[0235] ;

[0236] The constraint conditions are:

[0237] ;

[0238] ;

[0239] ;

[0240] Where is the value of the th input indicator of the th DMU, is the value of the th output indicator of the th DMU, and are the weights of the output and input indicators respectively, is a non-Archimedean infinitesimal, usually taking .

[0241] Solve this model through linear programming method to obtain the super-efficiency value of the th DMU.

[0242] Evaluate the relative efficiency of each maintenance plan in terms of economic benefits and social benefits. A super-efficiency value greater than 1 indicates that the plan is more optimal in terms of input-output.

[0243] Cover and Table of Contents: Include information such as the report name, preparation unit, date, etc., and list the table of contents of each part of the report.

[0244] Project Overview: Introduce the background, objectives and main maintenance plans of the maintenance project.

[0245] Pavement Performance Improvement Evaluation: List in detail the pavement performance improvement amplitude index of each maintenance plan, sort the plans, and analyze the key factors affecting the pavement performance improvement.

[0246] Economic and Social Benefit Evaluation: Present the super-efficiency value of each maintenance plan , analyze the advantages and disadvantages of each plan in terms of input and output, and point out the main ways to improve economic and social benefits.

[0247] Comprehensive evaluation and suggestions: Considering the improvement of pavement performance and the evaluation results of economic and social benefits, comprehensively rank each maintenance plan, give suggestions on the optimal maintenance plan, and propose improvement measures for the problems existing in each plan.

[0248] Example 7

[0249] Further illustrate in combination with Examples 1 - 6, integrate the highway technical condition assessment report, pavement performance prediction curve, maintenance demand list and fund allocation plan in steps S2 - S5; convert the road section geographical location information into geographical coordinates suitable for WebGL rendering;

[0250] Use the WebGL library to create a 3D scene and load the geographical base map; create a 3D model of the road section according to the road section geographical location, and use the formula Mark the disease location with a specific geometric shape;

[0251] Where is the base size, is the standardized disease degree data, is the scaling factor, used to control the change range of the disease mark size;

[0252] According to the maintenance demand list and fund allocation plan, represent the maintenance measures with different color lines or textures, and use the formula Show the maintenance progress in animation;

[0253] Where and are the starting color and ending color respectively, progress is the completion progress of the maintenance plan;

[0254] Bind the integrated data with Echarts, create bar charts, line charts, pie charts; add interactive events and query input boxes to the charts to achieve interactive query; provide summation and average statistical analysis functions, and display the statistical results with new charts or tables; set the data update period to achieve dynamic data update;

[0255] Use the formula Calculate the comprehensive benefit index;

[0256] Where , and are the weight coefficients, determined by expert scoring or historical data statistics, is the improvement value of the pavement performance PQI, and cost is the maintenance cost, is the quantitative value of social benefits;

[0257] Provide decision-making suggestions such as maintenance plan adjustment and fund allocation optimization for decision-makers according to the indicators.

[0258] The specific implementation method is as follows:

[0259] Collect the highway technical condition assessment report in step S2, the pavement performance prediction curve in step S3, the maintenance demand list in step S4, and the fund allocation plan in step S5. Integrate and correlate these data according to key information such as section number and geographical location to construct a unified data set.

[0260] Standardize the data such as pavement performance indicators (such as PQI, MQI) and disease severity so that their values are in the range of [0, 1] for subsequent visual display.

[0261] For the geographical location information of the section, convert it into geographical coordinates (such as longitude and latitude) suitable for WebGL rendering. If the original data is relative position information, it can be converted through relevant algorithms of the Geographic Information System (GIS).

[0262] Use a WebGL-related library (Three.js) to create a 3D scene. Define basic parameters such as the size, background color, and lighting of the scene.

[0263] Load the geographic base map, which can use open-source map data (OpenStreetMap), and convert it into a 3D terrain model as the basis for displaying the entire road condition.

[0264] Create a 3D model of the section in the 3D scene according to the geographical location information of the section. Cylinders or cuboids can be used to represent different types of sections, and their length, width, and height can be set according to the actual size of the section.

[0265] For the disease location, mark it with a specific geometric shape on the section model. The size and color of the disease can be set according to the severity of the disease (through the standardized disease severity data). For example, the more severe the disease, the larger the marked sphere and the brighter the color.

[0266] Disease marker size calculation formula:

[0267] ;

[0268] Where is the base size, is the standardized disease severity data, is the scaling factor, which is used to control the variation range of the size of disease markers. The function of this formula is to visually display the severity of the disease.

[0269] According to the maintenance requirement list and the fund allocation plan, visually display the maintenance plan for each section. Different colored lines or textures can be used to represent different types of maintenance measures (such as light maintenance, medium maintenance, heavy maintenance).

[0270] The time progress of the maintenance plan can be displayed through animation effects. Use color gradients or the dynamic growth of lines to represent the implementation progress of the maintenance plan.

[0271] Maintenance progress color gradient formula:

[0272] ;

[0273] where and are the starting color and the ending color respectively, and progress is the completion progress of the maintenance plan (the value range is between [0, 1]). This formula is used to dynamically display the implementation of the maintenance plan.

[0274] Bind the integrated data to Echarts, and create various charts according to different analysis requirements, such as bar charts (used to display the comparison of pavement performance indicators of different sections), line charts (used to display the pavement performance prediction curve), pie charts (used to display the fund allocation ratio), etc.

[0275] Add interactive events to the charts, such as mouse hovering, clicking, etc. When the user's mouse hovers over a chart element, display the detailed data information corresponding to that element, such as section number, specific pavement performance indicator values, maintenance plan details, etc.

[0276] Implement a query input box, and the user can enter keywords (such as section number, maintenance type, etc.) to filter and query data. The filtered data is updated in the chart display in real time.

[0277] Provide various statistical analysis functions, such as summation, average value, maximum value, minimum value, etc. The user can select different indicators and analysis dimensions (such as by section, by time, etc.) for statistical analysis.

[0278] The statistical analysis results are displayed through new charts or tables. For example, to statistically analyze the total fund investment in different maintenance types, use a bar chart to display the results.

[0279] Dynamic update mechanism: Set the data update period, and regularly obtain the latest data from the data source. When new data is updated, automatically refresh the display content in the Echarts chart and the WebGL 3D scene to ensure the real-time nature of the data.

[0280] Based on the results of visual display and statistical analysis, calculate some decision-related metrics. For example, calculate the comprehensive benefit metric for each road section, taking into account factors such as pavement performance improvement, maintenance cost, and social benefits.

[0281] Comprehensive benefit metric calculation formula:

[0282] ;

[0283] Where , and are weight coefficients, determined by expert scoring or historical data statistics, is the improvement value of pavement performance PQI, cost is the maintenance cost, is the quantified value of social benefits. This formula is used to comprehensively evaluate the maintenance benefits of each road section and provide a basis for decision-making.

[0284] Based on the calculated decision metrics, provide suggestions for decision-makers on aspects such as adjusting the maintenance plan and optimizing the fund allocation. For example, for road sections with higher comprehensive benefit metrics, the investment in maintenance funds can be appropriately increased; for road sections with lower comprehensive benefit metrics, re-evaluate the rationality of the maintenance plan.

[0285] Example 8

[0286] Further illustrate in combination with Example 1. As Figures 1-7 shown in the structure, it includes a data collection and preprocessing module, a road automatic detection system, a pavement performance prediction module, a maintenance decision module, a maintenance benefit evaluation module, and a visualization and analysis module;

[0287] The data collection and preprocessing module collects multi-source data of basic data, auxiliary basic data, and maintenance decision data on the public network and the highway management intranet, and uses a density-based spatial clustering algorithm for cleaning and integration, and a locally linear embedding algorithm for data dimensionality reduction processing;

[0288] The road automatic detection system is built using the Django framework based on Python, for functions such as data collection, index calculation, report generation and display, system testing and deployment, and uses the rough set theory algorithm to determine the detection index weights and calculate the PQI and MQI values;

[0289] The pavement performance prediction module is used to integrate relevant data, process outliers and construct derivative features, uses a composite kernel function and a Gaussian process regression model, and combines with the Bayesian optimization algorithm to predict pavement performance and generate a pavement performance prediction curve for the next 1 - 5 years;

[0290] The maintenance decision-making module is used to determine the maintenance priorities, predict the changing trend of maintenance requirements to generate a maintenance requirements list, adopt the cuckoo search algorithm to generate a fund allocation plan, and determine the maintenance plan, implementation cost, and expected effect data;

[0291] The maintenance benefit evaluation module uses the grey ideal solution method and the super-efficiency DEA model to evaluate the improvement amplitude of pavement performance, economic benefits, and social benefits before and after maintenance, and generates a maintenance benefit analysis report;

[0292] The visualization and analysis module uses 3D visualization technology based on WebGL to display road condition distribution, disease locations, and maintenance plan information, and uses the interactive query and statistical analysis functions based on Echarts to support dynamic data updates and calculate comprehensive benefit indicators to assist decision-making.

[0293] The above embodiments are only the preferred technical solutions of the present invention and should not be regarded as limitations on the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A scientific decision-making method for intelligent maintenance of highway pavements, characterized by: The method includes: S1. Collect multi-source data of basic data, auxiliary basic data, and maintenance decision data in the public network and the highway management intranet; preprocess the above multi-source data. S2. Build a road automatic detection system, input road damage, flatness, rut, and transverse force coefficient in the sections that need maintenance into the road automatic detection system, and the road automatic detection system generates a highway technical condition assessment report by combining PQI and MQI values. S3. Based on the highway technical condition assessment report in step S2, combined with historical detection data, traffic volume data, and climate data, select a Gaussian process regression model to predict pavement performance and generate a pavement performance prediction curve for the next 1 - 5 years. S4. Obtain maintenance standards and capital budget data according to the pavement performance prediction curve in step S3, and use ANP to determine the maintenance priorities of different sections for the maintenance standards and capital budget data. Combined with the priorities determined by ANP, apply the hidden Markov model to predict the change trend of maintenance requirements and generate a maintenance requirement list. S5. According to the maintenance requirement list in step S4, use the cuckoo search algorithm to generate a fund allocation plan and obtain maintenance plan, implementation cost, and expected effect data. S6. Input the maintenance plan, implementation cost, and expected effect data into the analysis system, evaluate the improvement range of pavement performance before and after maintenance, the economic and social benefits of maintenance, and generate a maintenance benefit analysis report. S7. Use the 3D visualization technology based on WebGL to display road condition distribution, disease location, and maintenance plan information for the highway technical condition assessment report, pavement performance prediction curve, maintenance requirement list, and fund allocation plan in S2 - S5. Then use the interactive query and statistical analysis function based on Echarts.

2. The scientific decision-making method for intelligent maintenance of a highway pavement according to claim 1, characterized in that: Use the density-based spatial clustering algorithm to clean and integrate the multi-source data and remove noise data. Adopt the locally linear embedding algorithm for data dimensionality reduction processing. Among them, the basic data includes: map GIS data, road network information data, highway pavement and subgrade data, and pavement technical condition assessment data. The auxiliary basic data includes: traffic flow data, meteorological data, geological environment data, and administrative management data. The maintenance decision data: maintenance material category data, maintenance equipment data, maintenance method data, and maintenance project unit price data.

3. The scientific decision-making method for intelligent maintenance of highway pavement according to claim 1, characterized in that: Build a road automatic detection system using the Django framework based on Python. Use the Django - admin startproject command to create a Django project, and then execute the python manage.pystartapp command to create a specific application. Configure the database connection information in the project's configuration file and select either MySQL or SQLite database; define the database models for road detection data and index calculation. The database models include road damage, evenness, rutting, and transverse force coefficient detection data, as well as PQI and MQI index data. Execute the commands "python manage.py makemigrations" and "python manage.py migrate" to create the corresponding database tables for the database models; Write a Python script to collect road damage, evenness, rutting, and transverse force coefficient data for the sections requiring maintenance; develop API interfaces using Django's REST framework to receive the collected data and store it in the database; Write Python code in a Django application to implement a rough set theory algorithm and construct an information system from the collected data , where is the set of objects, , where is the set of conditional attributes, i.e., each detection index, is the set of decision attributes, i.e., the overall performance of the road, is the set of attribute values, is the information function; Perform attribute reduction by calculating the discernibility matrix. The formula is: （ ）； Calculating Attribute Importance , where , ; Finally, according to determine the weights of each index; Standardize the original index data of PCI, RQI, RDI, and PBI that constitute PQI again for the collected data; Among them, PCI is the pavement condition index, RQI is the pavement ride quality index, RDI is the pavement rut depth index, and PBI is the pavement skid resistance performance index; Then, according to calculate the PQI value; For MQI, the original data of its constituent indicators PQI, BCI, TCI, and SCI are subjected to the same standardization process. The rough set theory is applied again to determine the weights of each indicator, and then according to calculate the MQI value; Among them, PQI is the pavement performance index, and MQI is the highway technical condition index; Based on the calculated PQI and MQI values and combined with the pre-set evaluation grade standards, generate a highway technical condition evaluation report including the basic information of the section, index calculation results, evaluation grade, and analysis and maintenance suggestions in the Django application; Use the Matplotlib and Echarts libraries to visually display the evaluation results; develop a Web interface using the Django template system to view the data and evaluation report information; Use the Django test framework to test each functional module of the system to ensure the normal functioning of data collection, calculation, and report generation functions; Deploy the system to a cloud server and use Nginx as the Web server and Gunicorn as the application server.

4. The scientific decision-making method for intelligent maintenance of highway pavement according to claim 1 is characterized in that: The specific steps of step S3 are: S31. Collect the PQI, MQI and each sub-item index value in the highway technical condition assessment report of step S2, and integrate them with the collected historical detection data, traffic volume data and climate data; use the method based on the local outlier factor to detect outliers. For the sample points in the dataset , calculate the local reachability density through the formula , calculate the local outlier factor through the formula . When is greater than the set threshold, process the outliers; Among them is of the set of nearest neighbor samples, , is the sample to its nearest neighbor distance, is the Euclidean distance between the sample and ; S32. Adopt a composite kernel function , for the training data set , calculate the covariance matrix , whose elements , considering the observation noise to obtain the covariance matrix with noise ; where and are the signal variances of the RBF and RQ kernel functions, respectively, and are their length scales, respectively, is the shape parameter of the rational quadratic kernel function, is the observation noise variance; S33. Define the log-likelihood function of Gaussian process regression , and use the Bayesian optimization algorithm to find the optimal hyperparameters ; wherein is the observation value vector of the training data, is the hyperparameter vector of the model, is the determinant of; S34. Generate prediction input data based on the time span of the next 1 - 5 years and the historical data change trend , calculate the covariance matrix between the prediction input and the training data and the covariance matrix of the prediction input itself , through the formula calculate the prediction mean, and through the formula calculate the prediction covariance; S35. Based on the prediction covariance matrix Calculate the confidence interval of the predicted value, and plot the changes of the predicted mean and the confidence interval over time to form a pavement performance prediction curve for the next 1 - 5 years.

5. The scientific decision-making method for intelligent maintenance of highway pavement according to claim 4, characterized in that: The specific steps of step S4 are: S41. Collect the pavement performance prediction curve data, maintenance standards, and capital budget data of step S3; use the method based on the interquartile range to process outliers, and determine the outlier judgment interval through the formula to determine the outlier judgment interval , and correct the outliers using the linear interpolation method; Use the min - max normalization formula to normalize the pavement performance prediction curve data; Among them, is the lower quartile, is the upper quartile; S42. Construct a network structure model including pavement performance, traffic impact, and economic cost element sets, analyze the interdependencies among the elements; construct a judgment matrix through the expert scoring method ; Solve to obtain the eigenvector and combine it into an unweighted supermatrix ; Determine the correlation weight matrix of the element sets , and calculate the weighted supermatrix ; Perform power operation to obtain the limit supermatrix , and determine the maintenance priority; Among them, is the eigenvector, is the matrix 's largest eigenvalue; S43. Define the set of maintenance requirement states and the set of observations ; Initialize the state transition probability matrix , the observation probability matrix and the initial state probability vector ; Use the Baum - Welch algorithm to train the model and update the parameters through the forward probability recurrence formula and the backward probability recurrence formula ; Use the Viterbi algorithm to find the most likely state sequence through the recurrence formula ; Find the most likely state sequence; S44. Generate a maintenance demand list including road section information, maintenance time, maintenance demand status, and estimated required funds according to the priority determined by ANP and the trend predicted by the hidden Markov model in the order of road section numbers and time, and adjust and screen maintenance projects according to the fund budget formula ; Among them, is i the funds required for each maintenance of a section of road, and is the total budget for funds.

6. The scientific decision-making method for intelligent maintenance of highway pavement according to claim 5, characterized in that: The specific steps of step S5 are: S51. Extract the road section maintenance requirement information, estimated required funds and maintenance priority from the maintenance requirement list in step S4 , and clarify the total budget of funds and the annual fund allocation ratio ; Define the objective function , where: Define the maintenance benefit function as ; Define the maintenance cost function as: ; Set constraint conditions and priority penalty terms ; S52. In the steps of the cuckoo search algorithm, select the solution with the largest objective function value as the optimal solution; Generate a fund allocation plan based on the optimal solution, determine the maintenance sections and allocated funds, clarify the maintenance plan; and calculate the total implementation cost and the expected effect data ; Among them, is the expected benefit improvement value after the maintenance of the road section i , is a decision variable and is the maintenance priority 7. A scientific decision-making method for intelligent maintenance of highway pavement according to claim 6, characterized in that: step S6 Input the maintenance plan, implementation cost, and expected effect data into the analysis system. The analysis system uses the grey ideal solution method to evaluate the improvement amplitude of the pavement performance before and after maintenance; The analysis system then uses the super - efficiency DEA model to evaluate the economic and social benefits of the maintenance and generates a maintenance benefit analysis report.

8. The scientific decision-making method for intelligent maintenance of highway pavement according to claim 7, characterized in that: The specific steps of step S6 are: S61. Collect the maintenance plan, implementation cost, and expected effect data in steps S51 and S52 and input them into the analysis system; perform standardization processing on the original data matrix to eliminate the influence of dimensions; S62. Determine the standardized data matrix and calculate the pavement performance improvement amplitude index; S63. Use the super - efficiency DEA model to evaluate the economic and social benefits: Taking each maintenance plan as a decision-making unit, determine the input implementation cost and output economic and social benefit indicators; establish a super-efficiency DEA model, and the objective function is ; The constraint conditions are: , , , the super-efficiency value is obtained by solving ; where is the th input index value of the th DMU, is the th output index value of the th DMU, and are the weights of the output and input indices respectively, is a non - Archimedean infinitesimal, where DMU is a decision - making unit; S64. The report includes a cover and table of contents, project overview, evaluation of pavement performance improvement , evaluation of economic and social benefits , comprehensive evaluation and suggestions, and appendices.

9. The scientific decision-making method for intelligent maintenance of highway pavement according to claim 8, characterized in that: Integrate the highway technical condition assessment reports, pavement performance prediction curves, maintenance requirement lists, and fund allocation plans in steps S2 - S5; convert the road section geographical location information into geographical coordinates suitable for WebGL rendering; Create a 3D scene using the WebGL library and load a geographic base map; create a 3D model of the road section according to the geographical location of the road section, and use the formula Mark the disease location with a specific geometric shape; Among them is the base size, is the disease severity data after standardization, is the scaling factor, which is used to control the variation range of the disease marker size; According to the maintenance requirement list and the fund allocation plan, use different colored lines or textures to represent the maintenance measures, and through the formula Show the maintenance progress in animation; wherein and are the starting color and the ending color respectively, progress is the completion progress of the maintenance plan; Bind the integrated data with Echarts to create bar charts, line charts, and pie charts; add interactive events and query input boxes to the charts to achieve interactive queries; provide statistical analysis functions such as summation and average value, and display the statistical results in new charts or tables; set the data update period to achieve dynamic data update; Use the formula to calculate the comprehensive benefit index; Among them 、 and are weight coefficients, determined by expert scoring or historical data statistics, is the improvement value of the pavement performance PQI, cost is the maintenance cost, is the quantified value of social benefits; Provide decision-making suggestions for adjusting the maintenance plan and optimizing the fund allocation for decision-makers according to the PQI and MQI indicators.

10. A scientific decision-making method for intelligent maintenance of highway pavement according to any one of claims 1-8, characterized in that: The decision-making system of the decision-making method includes a data collection and preprocessing module, a road automatic detection system, a pavement performance prediction module, a maintenance decision-making module, a maintenance benefit evaluation module, and a visualization and analysis module; The data collection and preprocessing module collects multi-source data of basic data, auxiliary basic data, and maintenance decision-making data on the public network and the highway management intranet, cleans and integrates them using the density-based spatial clustering algorithm, and performs data dimensionality reduction processing using the locally linear embedding algorithm; The road automatic detection system is built using the Django framework based on Python, has functions of data collection, index calculation, report generation and display, system testing and deployment, and determines the detection index weights and calculates the PQI and MQI values using the rough set theory algorithm; The pavement performance prediction module is used to integrate relevant data, process outliers and construct derivative features, uses a composite kernel function and a Gaussian process regression model, and combines with the Bayesian optimization algorithm to predict the pavement performance and generate a pavement performance prediction curve for the next 1 - 5 years; The maintenance decision-making module is used to determine the maintenance priority, predict the change trend of maintenance requirements to generate a maintenance requirement list, uses the cuckoo search algorithm to generate a fund allocation plan, and determines the maintenance plan, implementation cost, and expected effect data; The maintenance benefit evaluation module uses the grey ideal solution method and the super-efficiency DEA model to evaluate the improvement amplitude, economic benefits, and social benefits of the pavement performance before and after maintenance, and generates a maintenance benefit analysis report; The visualization and analysis module uses 3D visualization technology based on WebGL to display the road condition distribution, disease locations, and maintenance plan information, uses the interactive query and statistical analysis functions based on Echarts, supports dynamic data update, and calculates the comprehensive benefit index to assist decision-making.

Images