A method for decongesting high-speed traffic during construction based on data prediction

Abstract

The present application relates to the technical field of data prediction, and more particularly to a construction period high-speed traffic dispersal method based on data prediction, which comprises: dividing the construction influence range into several sub-regions, collecting historical traffic data for scenario deduction, constructing a dispersal strategy and setting expected data, periodically detecting actual data during the construction process, calculating the matching degree of each sub-region and the comprehensive matching degree to determine whether the dispersal is qualified, and optimizing the deduction or decision process after executing emergency dispersal when it is unqualified. The present application realizes dynamic closed-loop control of high-speed traffic dispersal during construction, can timely find the deviation between the dispersal scheme and the actual traffic condition and make targeted adjustment, effectively improves the adaptability and accuracy of traffic dispersal, and reduces the interference of construction on high-speed traffic.

Description

Technical Field

[0001] This invention relates to the field of data prediction technology, and in particular to a method for managing high-speed traffic during construction based on data prediction. Background Technology

[0002] During highway reconstruction and expansion or other roadside construction projects, the construction area often requires the closure of some lanes or sections, leading to a decrease in road capacity and potentially causing traffic congestion, delays, or even safety accidents. Traditional traffic management solutions during construction rely heavily on human experience for design, typically with fixed plans developed before construction begins and implemented according to the pre-determined plan during construction. This approach lacks the ability to dynamically respond to actual traffic conditions.

[0003] With the development of intelligent transportation systems, some methods have begun to incorporate real-time traffic monitoring data to make simple adjustments to traffic diversion plans, such as adjusting traffic light timings or issuing guidance information based on current traffic flow changes. However, existing methods generally have the following shortcomings: First, the design and construction impact range of diversion plans are not closely integrated with historical traffic patterns, making it difficult to accurately predict the chain reaction of construction on the surrounding road network; second, there is a lack of a quantitative evaluation mechanism for the effectiveness of diversion plan implementation, making it difficult to promptly identify deviations between the plan and the actual situation; third, when deviations occur, it is impossible to automatically locate whether the problem lies in the early simulation stage or the decision-making stage, resulting in a lack of targeted optimization and low iteration efficiency.

[0004] Chinese Patent Application No. 202411153903.3 discloses a traffic diversion method and system during the construction of a rail transit station. The method includes: acquiring construction information during the construction of the rail transit station; setting a traffic diversion speed evaluation model and calculating a traffic diversion speed index based on the construction information, which is used to evaluate the diversion speed in case of an emergency, wherein when the traffic diversion speed index is greater than a preset diversion threshold, the construction area is adjusted until it is less than or equal to the preset diversion threshold; acquiring all paths actually selected by pedestrians and setting a path selection evaluation model to evaluate all paths actually selected by pedestrians, finding the path with the smallest evaluation index as the optimal path, and setting road signs so that pedestrians can be diverted according to the optimal path.

[0005] However, existing technologies still have the following problems: There is a lack of dynamic quantitative evaluation of the relocation effect and an automatic mechanism for locating the root causes of problems and for targeted optimization. Summary of the Invention

[0006] To address this, the present invention provides a data-based method for managing high-speed traffic during construction, which overcomes the lack of dynamic quantitative evaluation of traffic management effectiveness and automatic identification and targeted optimization mechanisms for the root causes of problems in existing technologies.

[0007] To achieve the above objectives, this invention provides a method for highway traffic diversion during construction based on data prediction. It includes: Step S1: Obtain construction engineering data and preliminarily determine the scope of construction impact. Divide the scope of construction impact into several sub-regions according to the preset spatial grid size, and collect historical traffic data in each sub-region. Step S2: Based on the historical traffic data, a prediction method is used to extrapolate the scenario without traffic diversion, and the extrapolated scenario is obtained. Step S3: Based on the simulated scenario, construct a relocation decision for each sub-region and determine the expected data after implementing the relocation decision in each sub-region; Step S4: Begin construction and simultaneously implement the relocation decisions for each sub-area; Step S5: Periodically monitor the actual data of each sub-area within the construction impact range; Step S6: For each sub-region, calculate the matching degree of the sub-region based on the matching of the expected data and the actual data, and then calculate the comprehensive matching degree of all sub-regions. Based on the comprehensive matching degree, determine whether the traffic diversion during the current construction period is qualified. Step S7: For the current inspection cycle where traffic diversion is deemed unqualified during the current construction period, emergency diversion shall be implemented, and after the emergency diversion is completed, the scenario simulation process and / or diversion decision construction process shall be optimized.

[0008] Furthermore, calculating the overall matching degree of all sub-regions includes: For each sub-region, the matching degree of the sub-region is calculated based on the matching between the expected traffic density and the actual traffic density of the sub-region, as well as the matching between the expected vehicle type distribution and the actual vehicle type distribution of the sub-region. The overall matching score is obtained by weighted summing of the matching scores of all sub-regions. The weighting coefficient of each sub-region is determined based on the distance between the sub-region and the construction area, the historical average traffic flow density of the sub-region, the influencing factor, and the entanglement factor. The influencing factor represents the number of other sub-regions that can be affected after congestion occurs in the sub-region, and the entanglement factor represents the number of other sub-regions that can affect the traffic flow density of the sub-region.

[0009] Furthermore, determining whether traffic diversion is qualified during the current construction period based on the comprehensive matching degree includes: Compare the calculated overall matching score with the expected overall matching score; If the overall matching degree is greater than or equal to the expected overall matching degree, then the traffic diversion during the current construction period is deemed qualified. If the overall matching degree is less than the expected overall matching degree, the traffic diversion during the current construction period is deemed unqualified, emergency diversion is implemented, and the reason for the unqualified traffic diversion during the current construction period is determined after the emergency diversion is completed.

[0010] Furthermore, the implementation of emergency evacuation includes: Sub-regions with a matching degree less than the expected matching degree are marked as sub-regions to be cleared; For each sub-region to be relocated, the relocation requirements of the sub-region to be relocated are determined based on the distance between the sub-region to be relocated and other sub-regions to be relocated, as well as the difference in matching degree between the sub-regions to be relocated; the relocation requirements are broken down into multiple sub-requirements; For each sub-area to be diverted, qualified surrounding sub-areas are matched according to its sub-demands, and traffic flow is diverted to different qualified surrounding sub-areas in batches.

[0011] Furthermore, before diverting traffic flow to different qualified surrounding sub-areas in batches, the process also includes: For each qualified sub-region with a matching degree greater than or equal to the expected matching degree, calculate the unpacking value of the qualified sub-region; The evacuation value is corrected based on the distance between the qualified sub-region and the sub-region to be evacuated; wherein, the greater the distance, the greater the correction magnitude of the evacuation value; If the corrected relief value is greater than or equal to the sub-demand of the sub-region to be relieved, then the qualified sub-region is determined to be able to relieve the sub-region to be relieved. If the corrected relocation value is less than the sub-requirement, then the qualified sub-region is determined not to be relocated from the sub-region to be relocated.

[0012] Furthermore, after the emergency evacuation is completed, determine the reasons for the current construction period's traffic diversion failures, including: For each sub-region to be relocated, obtain the distance between it and the nearest sub-region to be relocated, and record this distance as the shortest neighbor distance for that sub-region to be relocated; After obtaining the shortest neighbor distance for all sub-regions to be cleared, calculate the variance of each shortest neighbor distance; If the variance is less than the preset variance, the reason for the failure is determined to be in the scenario simulation process; If the variance is greater than or equal to the preset variance, the reason for the failure is determined to be in the process of constructing the decision-making process.

[0013] Furthermore, when the reason for non-compliance is determined to be in the process of constructing the relocation decision-making mechanism, the relocation decision-making process should be optimized. Calculate the relocation decision application rate, which is the ratio of the number of sub-regions that have applied the relocation decision to the total number of sub-regions. The expected data for each sub-region is adjusted based on the application rate of the relocation decision; wherein, the higher the application rate of the relocation decision, the greater the increase in the expected data.

[0014] Furthermore, optimizing the decision-making process for relocation also includes: Calculate the overall relocation ratio; for each sub-area to be relocated, calculate the ratio of the traffic density eliminated after the emergency relocation decision to the difference between the actual density and the expected density, and record this ratio as the relocation ratio. The proportion of each relocation is corrected using a specific coefficient, and the average of the corrected proportions is recorded as the comprehensive relocation proportion; the specific coefficient is determined based on the relevant parameters of each sub-region to be relocated. The expected data is revised based on the overall relocation ratio; wherein, the larger the overall relocation ratio, the smaller the increase in the expected data.

[0015] Furthermore, if the reason for non-compliance lies in the traffic diversion decision-making process and the traffic diversion decision-making process is optimized, and the traffic diversion during the current construction period is still deemed non-compliant, then the reason for non-compliance will be re-determined to lie in the scenario simulation process.

[0016] Furthermore, when the reason for non-compliance lies in the scenario simulation process, the scenario simulation process is optimized, including: The comprehensive deviation is calculated based on the simulated traffic flow density and vehicle type distribution in the simulation scenario, and the actual traffic flow density and vehicle type distribution obtained from actual detection. The sample size of the historical traffic data is adjusted based on the comprehensive deviation; wherein, the larger the comprehensive deviation, the greater the increase in the sample size of the historical traffic data.

[0017] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention divides the construction impact area into several sub-regions and collects historical traffic data. Based on the historical data, it performs scenario simulation to obtain the simulated traffic flow density and vehicle type distribution under the condition of no traffic diversion. Then, it constructs diversion decisions and determines the expected data for each sub-region. During the construction process, it periodically detects the actual data, calculates the weighted matching degree of each sub-region based on the matching degree of traffic flow density and the matching degree of vehicle type distribution, and then obtains the comprehensive matching degree and compares it with the expected matching degree to determine whether the diversion is qualified. If it is unqualified, emergency diversion is executed. Based on the variance of the shortest adjacent distance of the sub-region to be diverted, it automatically determines whether the reason for the unqualification lies in the scenario simulation process or the diversion decision construction process, and optimizes the simulation parameters or decision parameters accordingly. This realizes dynamic closed-loop evaluation and adaptive optimization of highway traffic diversion during construction, effectively improving the response speed and accuracy of traffic diversion schemes and significantly reducing the risk of congestion caused by scheme deviations.

[0018] Furthermore, this invention divides the construction impact area into several sub-regions and collects historical traffic data. Based on the historical data, scenario simulations are performed to obtain the simulated traffic flow density and vehicle type distribution under non-relief conditions. Then, relief decisions are constructed for each sub-region and expected data is determined. During construction, actual data is periodically monitored, and the weighted matching degree of each sub-region based on the matching degree of traffic flow density and vehicle type distribution is calculated to obtain the comprehensive matching degree. By comparing the comprehensive matching degree with the expected comprehensive matching degree preset according to traffic importance level using ROC curve analysis based on historical emergency relief case data, the current relief is automatically determined to be qualified. If it is unqualified, emergency relief is triggered to further determine the reasons for the unqualification. This realizes the quantitative evaluation and adaptive judgment of highway traffic relief during construction, improving the timeliness and accuracy of relief plan response.

[0019] Furthermore, this invention marks sub-regions with a matching degree less than the expected matching degree as sub-regions to be alleviated, uses a dynamic adaptive method to determine the expected matching degree, calculates the alleviation demand based on the distance between the sub-regions to be alleviated and the difference in matching degree, and breaks down the alleviation demand into multiple sub-demands. Then, it matches each sub-demand with surrounding qualified sub-regions, and alleviates traffic flow in batches to different surrounding qualified sub-regions. This achieves decentralized and refined alleviation of congested traffic flow, avoids secondary congestion caused by centralized alleviation, and effectively improves the efficiency of emergency alleviation and the overall traffic capacity of the road network.

[0020] Furthermore, this invention calculates the basic diversion value of qualified sub-regions and corrects it based on the distance between them and the sub-regions to be diverted. The greater the distance, the greater the correction. The corrected diversion value is compared with the sub-demand to determine whether diversion is possible. This fully utilizes the characteristics of natural diversion of traffic flow through the road network during long-distance transfer, avoiding diversion failure or secondary congestion caused by inaccurate assessment of target area capacity, and improving the accuracy of emergency diversion decisions and the overall traffic efficiency of the road network.

[0021] Furthermore, this invention calculates the variance of the shortest adjacent distance of each sub-region to be relocated and compares it with a preset variance. When the variance is less than the preset variance, the reason for non-compliance is determined to be in the scenario simulation process; when the variance is greater than or equal to the preset variance, the reason for non-compliance is determined to be in the relocation decision-making process. This achieves automatic quantitative identification of the reasons for non-compliance, accurately distinguishing whether the problem lies in the early traffic situation simulation stage or the relocation decision-making stage. This provides a clear basis for subsequent targeted optimization, avoids blind adjustments, and improves optimization efficiency and accuracy.

[0022] Furthermore, this invention calculates the relocation decision application rate, which is the ratio of the number of sub-regions that actually implemented the relocation decision to the total number of sub-regions, and adjusts the expected data of each sub-region according to this application rate. The higher the relocation decision application rate, the greater the increase in expected data. This achieves adaptive optimization of the relocation decision construction process, enabling the expected data to dynamically reflect the actual implementation coverage of the relocation decision. This avoids frequent misjudgments caused by the disconnect between expected standards and implementation, and improves the rationality of the relocation effect evaluation and the accuracy of subsequent cycle determination.

[0023] Furthermore, this invention calculates the relocation ratio of each sub-region to be relocated, i.e., the ratio of the eliminated traffic density after the emergency relocation decision to the difference between the actual density and the expected density. After correcting each relocation ratio using specific coefficients based on relevant parameters of the sub-region, the average value is taken to obtain the comprehensive relocation ratio. Based on the comprehensive relocation ratio, the expected data is corrected in a negative correlation manner. The larger the comprehensive relocation ratio, the smaller the increase in the expected data. This achieves adaptive adjustment of the expected data based on the actual effect of emergency relocation, enabling the expected standard to dynamically match the real effectiveness of the relocation measures. This avoids evaluation distortion caused by standards that are too high or too low, and improves the scientific nature of the relocation effect evaluation and the rationality of subsequent decisions.

[0024] Furthermore, this invention, when the traffic diversion is still unqualified after optimization in the decision-making process, re-determines the cause as being in the scenario simulation process, and calculates the comprehensive deviation between the simulation data and the actual data. Based on the comprehensive deviation, the sample size of historical traffic data is corrected in a positive correlation manner. The larger the comprehensive deviation, the greater the increase in the sample size. This achieves closed-loop adaptive optimization of the scenario simulation process, enabling the simulation model to dynamically expand the training samples based on actual feedback, gradually improving the simulation accuracy. This fundamentally solves the problem of traffic diversion failure caused by prediction deviation, and improves the overall robustness and long-term adaptability of the high-speed traffic diversion method during construction. Attached Figure Description

[0025] Figure 1 This is a flowchart of the high-speed traffic diversion method during construction based on data prediction according to the present invention; Figure 2 This is a flowchart for determining whether the traffic diversion during the current construction period is qualified in the data prediction-based high-speed traffic diversion method of the present invention. Detailed Implementation

[0026] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0027] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0028] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0029] Please see Figures 1-2 As shown, Figure 1 This is a flowchart of the high-speed traffic diversion method during construction based on data prediction according to the present invention; Figure 2 This is a flowchart for determining whether the traffic diversion during the current construction period is qualified in the data prediction-based high-speed traffic diversion method of the present invention.

[0030] The present invention provides a data-predicted method for managing highway traffic during construction, comprising: Step S1: Obtain construction project data and preliminarily determine the scope of construction impact. Divide the scope of construction impact into several sub-regions according to the preset spatial grid size, and collect historical traffic data in each sub-region.

[0031] In this embodiment of the invention, construction engineering data is acquired, including the design scheme and implementation plan of the construction contract section. Based on the construction engineering data, the spatial range of the impact of construction on surrounding road traffic is preliminarily determined as the construction impact range. The construction impact range is divided into several sub-regions according to a preset spatial grid size. The preset spatial grid size is dynamically set according to the road grade and intersection density: a smaller grid size is used for main roads or areas with dense intersections; a larger grid size is used for secondary roads or areas with sparse intersections. Within each sub-region, historical traffic data is collected according to a preset time interval. The preset time interval is determined by using a clustering algorithm based on the hourly distribution characteristics of historical traffic flow to merge periods with small traffic flow fluctuations and subdivide periods with drastic traffic flow changes, automatically determining the boundaries of each time interval. The historical traffic data includes historical vehicle flow density and historical vehicle type distribution.

[0032] Step S2: Based on the historical traffic data, a prediction method is used to extrapolate the scenario without traffic diversion, and the extrapolated scenario is obtained.

[0033] In this embodiment of the invention, based on the historical traffic data, a time series prediction method is used to extrapolate the traffic conditions within the construction impact area without any traffic diversion intervention, resulting in a predicted scenario. The predicted scenario includes the predicted traffic flow density and predicted vehicle type distribution for each sub-region during the construction period. The parameters of the prediction method are determined using cross-validation with historical data.

[0034] Step S3: Based on the simulation scenario, construct a relocation decision for each sub-region and determine the expected data after implementing the relocation decision for each sub-region.

[0035] In this embodiment of the invention, based on the simulated scenario, and using the simulated traffic flow density and simulated vehicle type distribution of each sub-region as input, a traffic diversion decision is constructed for each sub-region. The diversion decision includes one or more combinations of lane reallocation, signal timing adjustment, and diversion path setting. Simultaneously, the expected data after implementing the diversion decision in each sub-region is determined. The expected data includes the expected traffic flow density and the expected vehicle type distribution. The expected data is determined by reducing the traffic flow density according to a preset diversion efficiency coefficient based on the simulated scenario. The diversion efficiency coefficient is obtained through statistical regression of historical diversion case data.

[0036] Step S4: Begin construction and simultaneously implement the relocation decisions for each sub-area.

[0037] In this embodiment of the invention, construction begins according to the construction organization and implementation plan, and at the same time as construction starts, the decongestion decisions for each sub-area constructed in step S3 are issued and executed.

[0038] Step S5: Periodically check the actual data of each sub-area within the construction impact range.

[0039] In this embodiment of the invention, during construction, actual data for each sub-area within the construction impact area is periodically monitored using roadside monitoring equipment according to a preset monitoring cycle. The preset monitoring cycle is determined based on the degree of historical traffic flow variation within the construction impact area; a smaller cycle value is used when changes are drastic, and a larger cycle value is used when changes are gradual. The actual data includes actual vehicle flow density and actual vehicle type distribution.

[0040] Step S6: For each sub-region, calculate the matching degree of the sub-region based on the matching of the expected data and the actual data, and then calculate the comprehensive matching degree of all sub-regions. Based on the comprehensive matching degree, determine whether the traffic diversion during the current construction period is qualified. Step S7: For the current inspection cycle where traffic diversion is deemed unqualified during the current construction period, emergency diversion shall be implemented, and after the emergency diversion is completed, the scenario simulation process and / or diversion decision construction process shall be optimized.

[0041] Specifically, calculating the overall matching degree of all sub-regions includes: For each sub-region, the matching degree of the sub-region is calculated based on the matching between the expected traffic density and the actual traffic density of the sub-region, as well as the matching between the expected vehicle type distribution and the actual vehicle type distribution of the sub-region. The overall matching score is obtained by weighted summing of the matching scores of all sub-regions. The weighting coefficient of each sub-region is determined based on the distance between the sub-region and the construction area, the historical average traffic flow density of the sub-region, the influencing factor, and the entanglement factor. The influencing factor represents the number of other sub-regions that can be affected after congestion occurs in the sub-region, and the entanglement factor represents the number of other sub-regions that can affect the traffic flow density of the sub-region.

[0042] Specifically, in this embodiment, the influencing factor is determined in the following way: starting from the sub-region, along the traffic flow direction, a breadth-first search is performed based on the road network topology until the correlation coefficient of the traffic flow density of the searched sub-regions is lower than a preset threshold, and the total number of searched sub-regions is the influencing factor; the related factor is determined in the following way: starting from the sub-region, the same search process is performed in the opposite direction of the traffic flow.

[0043] In this embodiment of the invention, firstly, for each sub-region i, the matching degree of traffic flow density and the matching degree of vehicle type distribution are calculated. The matching degree of traffic flow density is calculated as follows: the absolute difference between the expected traffic flow density and the actual traffic flow density is calculated, and the ratio of this absolute difference to the expected traffic flow density is taken as the relative error. This relative error is subtracted from 1 and multiplied by a first weighting coefficient to obtain the matching degree of traffic flow density. The matching degree of vehicle type distribution is calculated as follows: the expected vehicle type distribution and the actual vehicle type distribution are constructed as distribution vectors, and the cosine similarity between the two vectors is calculated. This cosine similarity is multiplied by a second weighting coefficient to obtain the matching degree of vehicle type distribution. The matching degree of traffic flow density and the matching degree of vehicle type distribution are added together to obtain the matching degree of the sub-region. The first and second weighting coefficients are determined by using principal component analysis or analytic hierarchy process (AHP) based on the contribution of traffic flow density data and vehicle type distribution data to traffic congestion in historical construction cases.

[0044] In this embodiment of the invention, the matching degrees of all sub-regions are weighted and summed to obtain the comprehensive matching degree. The weight coefficient w_i of each sub-region is determined as follows: A distance weight component is calculated based on the distance between the sub-region and the construction area; the closer the distance, the larger the distance weight component. Specifically, an inverse distance weighting method can be used, where the distance weight component is proportional to the reciprocal of the distance. A historical density weight component is calculated based on the historical average traffic density of the sub-region; the larger the historical average traffic density, the greater the contribution of the sub-region to the overall traffic situation, and the larger the historical density weight component. An influence weight component is calculated based on an influence factor, which represents the number of other sub-regions that can be affected by congestion in the sub-region. The larger the influence factor, the more significant the diffusion effect of congestion in the sub-region on the surrounding road network, and the larger the influence weight component. A connection weight component is calculated based on a connection factor, which represents the number of other sub-regions that can affect the traffic density of the sub-region. The larger the connection factor, the stronger the constraint of the surrounding road network on the traffic situation of the sub-region, and the larger the connection weight component.

[0045] In this embodiment of the invention, the distance weight component, historical density weight component, influence weight component, and related weight component are normalized, and then weighted and summed according to the preset contribution of each component to obtain the final weight coefficient w_i for the sub-region. The preset contribution of each component is determined by the analytic hierarchy process (AHP) based on the specific road network structure of the construction impact area. Finally, the matching degree of each sub-region is multiplied by the corresponding weight coefficient w_i and then summed to obtain the comprehensive matching degree.

[0046] This invention divides the construction impact area into several sub-regions and collects historical traffic data. Based on the historical data, scenario simulations are performed to obtain the simulated traffic flow density and vehicle type distribution under non-relief conditions. Then, relief decisions are constructed for each sub-region, and expected data are determined. During construction, actual data is periodically monitored, and the weighted matching degree of each sub-region based on the matching degree of traffic flow density and vehicle type distribution is calculated. The comprehensive matching degree is then obtained and compared with the expected matching degree to determine whether the relief is qualified. If it is unqualified, emergency relief is implemented. Based on the variance of the shortest adjacent distance of the sub-region to be relieved, the invention automatically determines whether the reason for the unqualification lies in the scenario simulation process or the relief decision construction process, and optimizes the simulation parameters or decision parameters accordingly. This achieves dynamic closed-loop evaluation and adaptive optimization of highway traffic relief during construction, effectively improving the response speed and accuracy of traffic relief schemes and significantly reducing the risk of congestion caused by scheme deviations.

[0047] Specifically, determining whether traffic diversion during the current construction period is up to standard based on the comprehensive matching degree includes: Compare the calculated overall matching score with the expected overall matching score; If the overall matching degree is greater than or equal to the expected overall matching degree, then the traffic diversion during the current construction period is deemed qualified. If the overall matching degree is less than the expected overall matching degree, the traffic diversion during the current construction period is deemed unqualified, emergency diversion is implemented, and the reason for the unqualified traffic diversion during the current construction period is determined after the emergency diversion is completed.

[0048] Specifically, in this embodiment, the "sub-area to be evacuated" information used for analysis after the emergency evacuation is completed is a snapshot of data saved before the emergency evacuation, rather than real-time data after the evacuation.

[0049] In this embodiment of the invention, firstly, the expected comprehensive matching degree is determined. The expected comprehensive matching degree is determined as follows: comprehensive matching degree data from multiple emergency evacuations in historical construction cases are collected; the comprehensive matching degree data is marked according to whether an emergency evacuation was triggered; a dichotomy method or ROC curve analysis is used to select the critical value of comprehensive matching degree that best distinguishes between qualified and unqualified states as the expected comprehensive matching degree; or, it is pre-set according to the traffic importance level of the construction impact area, with higher traffic importance levels resulting in a larger expected comprehensive matching degree.

[0050] In this embodiment of the invention, the comprehensive matching degree of the current detection cycle calculated in step S6 is numerically compared with the expected comprehensive matching degree. If the comprehensive matching degree is greater than or equal to the expected comprehensive matching degree, it indicates that the overall matching degree between the expected data and the actual data of each sub-region is within an acceptable range. Therefore, the traffic diversion during the current construction period is deemed qualified, and the periodic detection in step S5 continues, awaiting data from the next detection cycle. If the comprehensive matching degree is less than the expected comprehensive matching degree, it indicates that the overall matching degree between the expected data and the actual data of each sub-region is below an acceptable range, and the existing diversion decision fails to effectively address the actual traffic situation. Therefore, the traffic diversion during the current construction period is deemed unqualified, triggering the emergency diversion process. The emergency diversion operation in step S7 is executed, and after the emergency diversion is completed, the specific reasons for the current unqualification are further determined so that the subsequent scenario simulation process or diversion decision construction process can be optimized in a targeted manner.

[0051] This invention divides the construction impact area into several sub-regions and collects historical traffic data. Based on the historical data, it performs scenario simulations to obtain the simulated traffic density and vehicle type distribution under non-diversion conditions. Then, it constructs diversion decisions and determines expected data for each sub-region. During construction, it periodically detects actual data and calculates the weighted matching degree of each sub-region based on the matching degree of traffic density and vehicle type distribution, thereby obtaining the comprehensive matching degree. By comparing the comprehensive matching degree with the expected comprehensive matching degree preset according to traffic importance level using ROC curve analysis based on historical emergency diversion case data, it automatically determines whether the current diversion is qualified. If it is unqualified, it triggers emergency diversion and further determines the reason for the unqualification. Thus, it realizes the quantitative evaluation and adaptive judgment of highway traffic diversion during construction, improving the timeliness and accuracy of diversion plan response.

[0052] Specifically, the implementation of emergency evacuation includes: Sub-regions with a matching degree less than the expected matching degree are marked as sub-regions to be cleared; For each sub-region to be relocated, the relocation requirements of the sub-region to be relocated are determined based on the distance between the sub-region to be relocated and other sub-regions to be relocated, as well as the difference in matching degree between the sub-regions to be relocated; the relocation requirements are broken down into multiple sub-requirements; For each sub-area to be diverted, qualified surrounding sub-areas are matched according to its sub-demands, and traffic flow is diverted to different qualified surrounding sub-areas in batches.

[0053] In this embodiment of the invention, firstly, sub-regions with a matching degree less than the expected matching degree are marked as sub-regions to be relocated. Specifically, all sub-regions are traversed, and the current matching degree of each sub-region is compared with the expected matching degree. If the matching degree of a sub-region is lower than the expected matching degree, then the sub-region is marked as a sub-region to be relocated. The expected matching degree is determined using a dynamic adaptive method: the comprehensive matching degree of the previous detection cycle is smoothed according to a preset attenuation coefficient and then used as the expected matching degree of the current detection cycle. The attenuation coefficient is dynamically adjusted according to the rate of change of the construction progress.

[0054] Secondly, for each sub-area to be alleviated, the alleviation demand for that sub-area is determined based on the distance between it and other sub-areas to be alleviated, as well as the matching degree difference. The matching degree difference is the difference between the expected matching degree and the current matching degree of the sub-area; the larger the matching degree difference, the more severe the congestion in the sub-area, and the greater the traffic flow that needs to be alleviated. The distance is the spatial distance between the sub-area to be alleviated and other sub-areas to be alleviated; the closer the distance, the more mutual influence exists between congested areas, and the more coordinated the alleviation demand needs to be. The specific calculation method for the alleviation demand is as follows: the matching degree difference is used as the basic alleviation volume, and then the basic alleviation volume is corrected based on the distance to other sub-areas to be alleviated. The closer the distance, the larger the correction coefficient, indicating that priority should be given to alleviating the mutual influence of adjacent congested areas. A specific calculation example of the alleviation demand is as follows: Assuming the current matching degree of a sub-region A to be relocated is 0.65, and the expected matching degree is 0.85, then the matching degree difference ΔM = 0.85 - 0.65 = 0.20. Using this matching degree difference as the base relocation volume, let the base relocation volume Q_base = ΔM × standard traffic flow conversion factor, where the standard traffic flow conversion factor is 1000 vehicles / unit matching degree, then Q_base = 0.20 × 1000 = 200 vehicles. Let another sub-region B adjacent to sub-region A be relocated, with a distance of d_AB = 0.5km between A and B. Set the distance correction factor k_distance = 1 + (d_ref / d_AB), where d_ref is the reference distance, taken as 1km, then k_distance = 1 + (1 / 0.5) = 1 + 2 = 3. Simultaneously, considering other sub-regions C and D around A, calculate the distance correction factors between A and C, and between A and D respectively, and take the maximum value among all correction factors as the comprehensive correction factor. Let the distance between A and C be 0.8km, with a correction factor of 1 + 1 / 0.8 = 2.25; the distance between A and D be 1.2km, with a correction factor of 1 + 1 / 1.2 ≈ 1.83; taking the maximum correction factor of 3. Then the relocation demand of sub-region A is Q_demand = Q_base × k_max = 200 × 3 = 600 vehicles. This relocation demand indicates that sub-region A needs to relocate 600 equivalent vehicles to surrounding qualified sub-regions.

[0055] In this embodiment of the invention, the calculated traffic diversion demand is divided into multiple sub-demands, where the value of each sub-demand is less than the total traffic flow that needs to be diverted from the sub-area to be diverted. The division method is either uniform division based on the number of surrounding qualified sub-areas, or non-uniform division based on the diversion capacity of each surrounding qualified sub-area. The diversion capacity is determined by the historical remaining capacity or real-time idle capacity of the qualified sub-areas. Finally, for each sub-area to be diverted, surrounding qualified sub-areas are matched according to each of its sub-demands, and the traffic flow is diverted in batches to different surrounding qualified sub-areas. Specifically, for each sub-demand, the surrounding qualified sub-areas are traversed, the remaining diversion capacity of each qualified sub-area is calculated, and qualified sub-areas with remaining diversion capacity greater than or equal to the sub-demand are selected as target diversion areas. The batch of traffic flow is then guided from the sub-area to be diverted to the target diversion area. If the remaining diversion capacity of the surrounding qualified sub-areas cannot meet a certain sub-demand, the sub-demand is further divided until a matching qualified sub-area is found. A specific embodiment of dividing the diversion demand into sub-demands and matching them with surrounding qualified sub-areas is as follows: Suppose that the relocation demand of a certain sub-region A is equivalent to 600 vehicles, and there are 3 qualified sub-regions B, C and D in the surrounding area.

[0056] A uniform splitting method is adopted: based on the number of three qualified sub-regions, the 600 vehicles are evenly divided into three sub-demands, each containing 200 vehicles. The dispersal capacity of each qualified sub-region is assessed: Sub-region B has a remaining dispersal capacity of 250 vehicles, which is greater than 200 vehicles, and can accommodate the first batch of 200 vehicles; Sub-region C has a remaining dispersal capacity of 180 vehicles, which is less than 200 vehicles, and cannot directly accommodate 200 vehicles; Sub-region D has a remaining dispersal capacity of 300 vehicles, which is greater than 200 vehicles, and can accommodate the third batch of 200 vehicles. For cases where sub-region C cannot accommodate the vehicles, the corresponding 200-vehicle sub-demand is further split into two sub-demands of 100 vehicles each, and then re-matched with qualified sub-regions.

[0057] A non-uniform allocation method is adopted: The vehicles are non-uniformly allocated based on the dispersal capacity of each qualified sub-region. Assume the real-time idle capacity of sub-region B is 250 vehicles, sub-region C is 150 vehicles, and sub-region D is 200 vehicles, with a total remaining dispersal capacity of 600 vehicles. The non-uniform allocation is performed according to the proportion of the remaining dispersal capacity of each sub-region: the allocation ratio for sub-region B is 250 / 600≈0.417, corresponding to a sub-demand of 600×0.417≈250 vehicles; the allocation ratio for sub-region C is 150 / 600=0.25, corresponding to a sub-demand of 150 vehicles; and the allocation ratio for sub-region D is 200 / 600≈0.333, corresponding to a sub-demand of 200 vehicles. Since each sub-demand is less than or equal to the remaining dispersal capacity of its corresponding qualified sub-region, no further allocation is needed, and the matching is successful directly.

[0058] Sub-demand matching process: Taking a uniform split as an example, for a sub-demand of 200 vehicles, the surrounding qualified sub-regions are traversed. First, sub-region B is evaluated. Its remaining capacity is 250 vehicles, which is greater than 200 vehicles, so the match is successful. The first batch of 200 vehicles is redirected from sub-region A to sub-region B, and the remaining capacity of sub-region B is reduced by 50 vehicles. For the sub-demand of 200 vehicles corresponding to sub-region C, the remaining capacity of sub-region C is evaluated. It is 180 vehicles, which is less than 200 vehicles, so the match fails. Sub-region D is then evaluated. Its remaining capacity is 300 vehicles, which is greater than 200 vehicles, so the match is successful, and the batch of traffic is redirected to sub-region D. If the remaining capacity of all surrounding qualified sub-regions is insufficient to meet a certain sub-demand, for example, if a certain sub-demand is 300 vehicles, and the remaining capacity of B, C, and D is 100, 80, and 120 vehicles respectively, all less than 300 vehicles, then the 300 vehicle sub-demand is further divided into three smaller sub-demands of 100, 100, and 100 vehicles, and B, C, and D are re-matched respectively, until each sub-demand has a corresponding qualified sub-region to accept it.

[0059] This invention marks sub-regions with a matching degree less than the expected matching degree as sub-regions to be alleviated, uses a dynamic adaptive method to determine the expected matching degree, calculates the alleviation demand based on the distance between the sub-regions to be alleviated and the difference in matching degree, and breaks down the alleviation demand into multiple sub-demands. Then, it matches each sub-demand with a qualified surrounding sub-region, and alleviates traffic flow in batches to different qualified surrounding sub-regions. This achieves decentralized and refined alleviation of congested traffic flow, avoids secondary congestion caused by centralized alleviation, and effectively improves the efficiency of emergency alleviation and the overall traffic capacity of the road network.

[0060] Specifically, before diverting traffic flow to different qualified surrounding sub-areas in batches, the process also includes: For each qualified sub-region with a matching degree greater than or equal to the expected matching degree, calculate the unpacking value of the qualified sub-region; The evacuation value is corrected based on the distance between the qualified sub-region and the sub-region to be evacuated; wherein, the greater the distance, the greater the correction magnitude of the evacuation value; If the corrected relief value is greater than or equal to the sub-demand of the sub-region to be relieved, then the qualified sub-region is determined to be able to relieve the sub-region to be relieved. If the corrected relocation value is less than the sub-requirement, then the qualified sub-region is determined not to be relocated from the sub-region to be relocated.

[0061] In this embodiment of the invention, firstly, for each qualified sub-region with a matching degree greater than or equal to the expected matching degree, the diversion value of the qualified sub-region is calculated. The diversion value characterizes the ability of the qualified sub-region to accept additional traffic flow. The diversion value is calculated as follows: the remaining road capacity of the qualified sub-region is obtained, which is obtained by subtracting the current actual traffic flow from the design capacity of the sub-region; the remaining road capacity is multiplied by a preset diversion coefficient to obtain the basic diversion value, which is determined based on the historical diversion efficiency statistics of the sub-region and ranges from 0 to 1.

[0062] Secondly, the diversion value is adjusted based on the distance between the qualified sub-region and the sub-region to be diverted. The greater the distance, the larger the adjustment range of the diversion value. The physical meaning is that as traffic flow from the sub-region to be diverted moves to a more distant qualified sub-region, the road network along the route naturally diverts some traffic flow, thus allowing the more distant target area to accommodate more diverted traffic. Specifically, a distance correction factor is calculated, which is proportional to the distance. The base diversion value is multiplied by this distance correction factor to obtain the corrected diversion value. The lower limit of the distance correction factor is 1, meaning the correction factor is 1 when the distance is zero; the larger the distance, the larger the correction factor value.

[0063] In this embodiment of the invention, the specific implementation of the qualified sub-area traffic diversion capacity determination and distance correction is as follows: Assume that a qualified sub-area B exists around a certain sub-area A to be diverted. The designed traffic capacity of sub-area B is 1200 vehicles / hour, and the current actual traffic flow is 800 vehicles / hour. Then, the remaining road capacity of sub-area B is 1200-800=400 vehicles / hour. The preset diversion coefficient is determined based on the historical diversion efficiency statistics of sub-area B. The ratio of the actual reception flow to the remaining capacity of sub-area B in the past 10 emergency diversions is collected, and the average value is taken as the diversion coefficient. Assuming that the average diversion efficiency obtained from historical statistics is 0.85, then the basic diversion value = remaining road capacity × diversion coefficient = 400 × 0.85 = 340 vehicles / hour. Assume that the distance between sub-area A and sub-area B is d=2km. The distance correction factor is calculated as: k_distance=1+α×d, where α is the distance correction coefficient, and α=0.3 / km. Therefore, k_distance = 1 + 0.3 × 2 = 1.6. The corrected diversion value = basic diversion value × k_distance = 340 × 1.6 = 544 vehicles / hour. The physical meaning of this correction is that during the transfer of traffic flow from sub-region A to sub-region B 2km away, approximately 60% of the flow will be naturally diverted through the road network. Therefore, sub-region B can actually accommodate 544 vehicles / hour of transferred traffic from A, which is higher than its basic diversion value of 340 vehicles / hour.

[0064] Suppose a sub-demand of sub-area A to be alleviated is 500 vehicles / hour. Compare the corrected alleviation value of 544 vehicles / hour with the sub-demand of 500 vehicles / hour: 544 ≥ 500, then sub-area B is deemed capable of alleviating this batch of traffic flow from sub-area A. If another sub-demand of sub-area A is 600 vehicles / hour, then 544 < 600, and sub-area B is deemed unable to alleviate this batch of traffic flow, requiring the search for other qualified sub-areas or further subdivision of the sub-demand. If the distance between sub-area A and sub-area B is d = 0.5 km, then k_distance = 1 + 0.3 × 0.5 = 1.15, and the corrected alleviation value = 340 × 1.15 = 391 vehicles / hour. When the distance is short, the correction is smaller, indicating limited natural diversion along the way, and the target area needs to handle more direct transfer traffic. If the distance d between sub-region A and sub-region B is 5km, then k_distance = 1 + 0.3 × 5 = 2.5, and the corrected diversion value is 340 × 2.5 = 850 vehicles / hour. A larger correction indicates a more significant natural diversion effect along the route, meaning the target area can accommodate more transferred traffic.

[0065] The revised diversion value is compared with the sub-demand of the sub-area to be diverted. If the revised diversion value is greater than or equal to the sub-demand, it indicates that the qualified sub-area has the capacity to accept the batch of traffic flow, and the qualified sub-area is determined to be able to divert the sub-area to be diverted, and is added to the candidate target area list. If the revised diversion value is less than the sub-demand, it indicates that the remaining capacity of the qualified sub-area is insufficient to accept the batch of traffic flow, and the qualified sub-area is determined not to be able to divert the sub-area to be diverted, and the next qualified sub-area is evaluated.

[0066] This invention calculates the basic diversion value of qualified sub-regions and corrects it based on the distance between them and the sub-regions to be diverted. The greater the distance, the greater the correction. The corrected diversion value is compared with the sub-demand to determine whether diversion is possible. This invention makes full use of the natural diversion characteristics of traffic flow through the road network during long-distance transfer, avoiding diversion failure or secondary congestion caused by inaccurate assessment of target area capacity, and improving the accuracy of emergency diversion decisions and the overall traffic efficiency of the road network.

[0067] Specifically, after the emergency evacuation is completed, the reasons for the current construction period's traffic management failures should be determined, including: For each sub-region to be relocated, obtain the distance between it and the nearest sub-region to be relocated, and record this distance as the shortest neighbor distance for that sub-region to be relocated; After obtaining the shortest neighbor distance for all sub-regions to be cleared, calculate the variance of each shortest neighbor distance; If the variance is less than the preset variance, the reason for the failure is determined to be in the scenario simulation process; If the variance is greater than or equal to the preset variance, the reason for the failure is determined to be in the process of constructing the decision-making process.

[0068] In this embodiment of the invention, firstly, for each sub-region to be relocated, the distance between it and the nearest sub-region to be relocated is obtained, and this distance is recorded as the shortest neighbor distance for that sub-region. Specifically, assuming there are m sub-regions to be relocated, for the i-th sub-region to be relocated, the Euclidean distance or road network distance between it and the remaining m-1 sub-regions to be relocated is calculated, and the minimum value is taken as the shortest neighbor distance d_i for that sub-region. For example, suppose there are 10 sub-regions within a certain construction impact area, and 5 of them are identified as sub-regions to be relocated, numbered R1, R2, R3, R4, and R5 respectively. The shortest adjacent distances of each sub-region are calculated as follows: the distance between R1 and the nearest sub-region to be relocated, R2, is 0.3km; the distance between R2 and the nearest sub-region to be relocated, R1, is 0.3km; the distance between R3 and the nearest sub-region to be relocated, R4, is 0.5km; the distance between R4 and the nearest sub-region to be relocated, R3, is 0.5km; and the distance between R5 and the nearest sub-region to be relocated, R4, is 1.2km.

[0069] Then, after obtaining the shortest neighbor distances for all sub-regions to be cleared, the variance of each shortest neighbor distance is calculated. Taking the above data as an example, the shortest neighbor distance sequence is [0.3, 0.3, 0.5, 0.5, 1.2]. The mean μ is calculated as (0.3 + 0.3 + 0.5 + 0.5 + 1.2) / 5 = 0.56 km, and the variance σ² is calculated as [(0.3 - 0.56)² + (0.3 - 0.56)² + (0.5 - 0.56)² + (0.5 - 0.56)² + (1.2 - 0.56)²] / 5 = (0.0676 + 0.0676 + 0.0036 + 0.0036 + 0.4096) / 5 = 0.552 / 5 = 0.1104. The preset variance is determined based on the statistical results of the classification of non-conformity reasons in historical construction cases. Variance data of the shortest adjacent distances for each sub-area to be relocated in multiple historical construction cases were collected, and the corresponding reasons for non-compliance (deviations in the scenario simulation process or deviations in the relocation decision-making process) were manually labeled. Cluster analysis or ROC curve analysis was used to determine the optimal variance threshold for distinguishing between the two types of reasons. The preset variance threshold was set to 0.08.

[0070] Comparing the calculated variance of 0.1104 with the preset variance of 0.08: If 0.1104 ≥ 0.08, the reason for disqualification lies in the construction process of the traffic diversion decision. The physical meaning is that the spatial distribution of the sub-regions to be diverted is relatively dispersed (large variance), indicating that congestion points are randomly distributed across various sub-regions within the affected area, rather than concentrated in a specific area. This usually indicates that the differences between sub-regions were not fully considered during the construction of the traffic diversion decision, resulting in poor diversion effects in some sub-regions, rather than a systematic bias in the prediction of the overall traffic situation during the simulation phase. If the calculated variance is less than the preset variance, for example, a variance of 0.05, the reason for disqualification lies in the scenario simulation process. The physical meaning is that the sub-regions to be diverted are concentrated in a certain local area (small variance), indicating a significant deviation between the actual traffic conditions in that area and the simulated scenario. The prediction of that area during the simulation process was inaccurate, causing the traffic diversion decision to fail in that area.

[0071] This invention calculates the variance of the shortest adjacent distance for each sub-region to be relocated and compares it with a preset variance. When the variance is less than the preset variance, the reason for non-compliance is determined to be in the scenario simulation process; when the variance is greater than or equal to the preset variance, the reason for non-compliance is determined to be in the relocation decision-making process. This achieves automatic quantitative identification of the reasons for non-compliance, accurately distinguishing whether the problem originates in the early traffic situation simulation stage or the relocation decision-making stage. This provides a clear basis for subsequent targeted optimization, avoids blind adjustments, and improves optimization efficiency and accuracy.

[0072] Specifically, when the reason for non-compliance lies in the decision-making process for relocation, the decision-making process for relocation should be optimized. Calculate the relocation decision application rate, which is the ratio of the number of sub-regions that have applied the relocation decision to the total number of sub-regions. The expected data for each sub-region is adjusted based on the application rate of the relocation decision; wherein, the higher the application rate of the relocation decision, the greater the increase in the expected data.

[0073] In this embodiment of the invention, firstly, the application rate of traffic diversion decisions is calculated. The application rate of traffic diversion decisions is the ratio of the number of sub-regions that applied the diversion decisions to the total number of sub-regions. Specifically, all sub-regions are traversed, and the number of sub-regions that actually implemented the diversion decisions constructed in step S3 during the actual construction process is counted, denoted as N_applied. The total number of sub-regions is N_total, then the application rate of traffic diversion decisions R_applied = N_applied / N_total. For example, if there are 10 sub-regions within the construction impact area, and 7 of them implemented corresponding traffic diversion measures (such as lane reallocation, signal timing adjustment, etc.) according to the pre-constructed diversion decisions during the actual construction process, while the other 3 sub-regions failed to implement the pre-constructed diversion decisions or implemented other temporary measures due to a large deviation between the actual traffic conditions and expectations, then the application rate of traffic diversion decisions R_applied = 7 / 10 = 0.7.

[0074] Secondly, the expected data for each sub-region is adjusted based on the application rate of the relocation decision. The expected data includes expected traffic density and expected vehicle type distribution. The adjustment method is as follows: the original expected data is multiplied by an adjustment coefficient, which is positively correlated with the application rate of the relocation decision; that is, the higher the application rate, the greater the increase in the expected data. The specific calculation method for the adjustment coefficient is: k_adjust=1+β×R_applied, where β is a preset adjustment intensity coefficient, ranging from 0 to 1, determined based on the historical data fluctuation of the construction impact area. Let β=0.5, then k_adjust=1+0.5×0.7=1.35. The original expected traffic density of each sub-region is multiplied by 1.35 to obtain the adjusted expected traffic density; similarly, each component of the expected vehicle type distribution is multiplied by the same adjustment coefficient and normalized to obtain the adjusted expected vehicle type distribution.

[0075] The physical significance of adjustment lies in the following: a higher application rate of traffic diversion decisions indicates a wider coverage of the pre-constructed diversion decisions being adopted and implemented in actual construction, and a higher overall credibility of the diversion decisions. In this case, the standard of expected data should be appropriately raised, requiring subsequent cycles to achieve higher diversion effects. Conversely, a lower application rate of diversion decisions indicates that a large number of sub-areas have not been implemented according to the preset diversion decisions, and the expected data needs to be lowered or kept unchanged to avoid frequent triggering of non-compliance judgments due to excessively high standards. After adjustment, the adjusted expected data will be used as the benchmark for the next monitoring cycle, and steps S5 to S6 will be repeated to continue monitoring the traffic diversion effect during construction.

[0076] This invention calculates the relocation decision application rate, which is the ratio of the number of sub-regions that actually implemented the relocation decision to the total number of sub-regions. Based on this application rate, the expected data for each sub-region is adjusted, and the higher the application rate, the greater the increase in expected data. This achieves adaptive optimization of the relocation decision construction process, enabling the expected data to dynamically reflect the actual implementation coverage of the relocation decision. This avoids frequent misjudgments caused by the disconnect between expected standards and implementation, and improves the rationality of the relocation effect evaluation and the accuracy of subsequent cycle determination.

[0077] Specifically, optimizing the decision-making process for relocation also includes: Calculate the overall relocation ratio; for each sub-area to be relocated, calculate the ratio of the traffic density eliminated after the emergency relocation decision to the difference between the actual density and the expected density, and record this ratio as the relocation ratio. The proportion of each relocation is corrected using a specific coefficient, and the average of the corrected proportions is recorded as the comprehensive relocation proportion; the specific coefficient is determined based on the relevant parameters of each sub-region to be relocated. The expected data is revised based on the overall relocation ratio; wherein, the larger the overall relocation ratio, the smaller the increase in the expected data.

[0078] In this embodiment of the invention, firstly, the overall traffic flow reduction ratio is calculated. For each sub-area to be reduced, the ratio of the traffic density eliminated after the emergency reduction decision to the difference between the actual density and the expected density is calculated, and this ratio is recorded as the reduction ratio. Specifically, let the actual traffic density of the i-th sub-area to be reduced be ρ_i_actual, the expected traffic density be ρ_i_expected, and after the implementation of the emergency reduction decision, the traffic density of this sub-area decreases to ρ_i_relieved. Then, the eliminated traffic density is ρ_i_actual - ρ_i_relieved, and the difference between the actual density and the expected density is ρ_i_actual - ρ_i_expected. The reduction ratio η_i = (ρ_i_actual - ρ_i_relieved) / (ρ_i_actual - ρ_i_expected). This ratio characterizes the efficiency of the emergency reduction measures in eliminating the actual excess density; the closer the ratio is to 1, the better the reduction effect.

[0079] Secondly, specific coefficients are used to correct the proportions of each relocation, and the average of the corrected proportions is recorded as the comprehensive relocation proportion. The specific coefficients are determined based on relevant parameters of each sub-region to be relocated. These relevant parameters include one or more combinations of the distance between the sub-region and the construction area, the historical average traffic density of the sub-region, influencing factors, and entanglement factors. The specific coefficients are determined as follows: for the i-th sub-region to be relocated, its relevant parameters are normalized and then weighted and summed to obtain the correction coefficient γ_i for that sub-region. The corrected relocation proportion is then η_i' = γ_i × η_i. The comprehensive relocation proportion η_total = (1 / m) × Ση_i', where m is the total number of sub-regions to be relocated.

[0080] Then, the expected data is corrected based on the comprehensive relocation ratio. The correction method is as follows: the expected data for the current monitoring cycle is multiplied by an expected adjustment coefficient. This coefficient is negatively correlated with the comprehensive relocation ratio; that is, the larger the comprehensive relocation ratio, the smaller the increase in the expected data. The specific calculation method for the expected adjustment coefficient is: k_expected = 1 - λ × η_total, where λ is a preset adjustment sensitivity coefficient, ranging from 0 to 1, determined based on the sensitivity of the construction impact area to the relocation effect. The expected traffic density and expected vehicle type distribution of each sub-region are multiplied by k_expected to obtain the corrected expected data. The physical meaning of this correction is: a larger comprehensive relocation ratio indicates that the current emergency relocation measures have achieved good results, and the actual congestion situation has been effectively alleviated. In this case, it is not necessary to excessively increase the expected standards for subsequent cycles, and even appropriately reduce the increase in expected data to maintain the stability of the relocation decision; conversely, a smaller comprehensive relocation ratio indicates that the emergency relocation effect is not ideal, and it is necessary to appropriately increase the expected data standards to make the relocation decisions for subsequent cycles more stringent and effective. After the correction is completed, the corrected expected data will be used as the benchmark for the next detection cycle, and steps S5 to S6 will be executed again.

[0081] This invention calculates the relocation ratio of each sub-region to be relocated, which is the ratio of the eliminated traffic density after the emergency relocation decision to the difference between the actual density and the expected density. After correcting each relocation ratio with specific coefficients based on relevant parameters of the sub-region, the average value is taken to obtain the comprehensive relocation ratio. Based on the comprehensive relocation ratio, the expected data is corrected in a negative correlation manner. The larger the comprehensive relocation ratio, the smaller the increase in the expected data. This realizes the adaptive adjustment of expected data based on the actual effect of emergency relocation, so that the expected standard can dynamically match the real effect of relocation measures, avoid the evaluation distortion caused by the standard being too high or too low, and improve the scientific nature of the relocation effect evaluation and the rationality of subsequent decisions.

[0082] Specifically, if the reason for non-compliance lies in the traffic diversion decision-making process and the traffic diversion decision-making process is optimized, and the traffic diversion during the current construction period is still deemed non-compliant, then the reason for non-compliance will be re-determined to lie in the scenario simulation process.

[0083] Specifically, when the reason for non-compliance lies in the scenario simulation process, the scenario simulation process is optimized, including: The comprehensive deviation is calculated based on the simulated traffic flow density and vehicle type distribution in the simulation scenario, and the actual traffic flow density and vehicle type distribution obtained from actual detection. The sample size of the historical traffic data is adjusted based on the comprehensive deviation; wherein, the larger the comprehensive deviation, the greater the increase in the sample size of the historical traffic data.

[0084] In this embodiment of the invention, firstly, after determining that the reason for non-compliance lies in the traffic diversion decision-making process and optimizing the traffic diversion decision-making process, if the traffic diversion is still determined to be non-compliant during the current construction period, then the reason for non-compliance is re-determined to be the scenario simulation process. Specifically, after performing the optimization operation described in claim 7 or claim 8, the next detection cycle begins, and steps S5 to S6 are re-executed. If, in the new detection cycle, the overall matching degree is still less than the expected matching degree, that is, the traffic diversion is still determined to be non-compliant, it indicates that the optimization of the traffic diversion decision-making process has failed to solve the problem, and the root cause of non-compliance lies in the prediction deviation of the scenario simulation process itself. At this time, the reason for non-compliance is re-determined to be the deviation of the scenario simulation process.

[0085] Secondly, when the reason for non-compliance lies in the scenario simulation process, the scenario simulation process is optimized. Based on the simulated traffic flow density and vehicle type distribution in the simulated scenario, and comparing them with the actual traffic flow density and vehicle type distribution obtained from actual detection, a comprehensive deviation is calculated. The comprehensive deviation is calculated as follows: for each sub-region, the relative error between the simulated and actual traffic flow density, and the divergence or cosine distance between the simulated and actual vehicle type distributions are calculated separately. These two values ​​are then weighted and summed to obtain the local deviation for that sub-region. The local deviations of all sub-regions are then weighted and averaged to obtain the comprehensive deviation. The weighting coefficients are determined based on factors such as the distance between each sub-region and the construction area, and the historical average traffic flow density.

[0086] Then, the sample size of the historical traffic data is adjusted based on the comprehensive deviation. The adjustment method is as follows: the current sample size of the historical traffic data used for scenario deduction in step S2 is multiplied by a sample size adjustment coefficient. This coefficient is positively correlated with the comprehensive deviation; that is, the larger the comprehensive deviation, the greater the increase in the sample size of the historical traffic data. The specific calculation method for the sample size adjustment coefficient is: k_sample = 1 + θ × δ, where δ is the comprehensive deviation and θ is a preset sample size adjustment intensity coefficient, determined based on the fluctuation degree of historical data and the complexity of the construction impact range. The original historical traffic data sample size is multiplied by k_sample to obtain the expanded sample size. Expansion methods include: supplementing data from earlier periods within the same time interval from the historical database, or generating synthetic samples through data augmentation methods (such as adding noise, time offset, etc.).

[0087] The physical significance of this optimization is that a larger overall deviation indicates a greater gap between the current simulated scenario and the actual traffic conditions. This means the simulation model's predictive ability for traffic conditions within the construction impact area is insufficient, necessitating an increase in the sample size of historical data to enrich the basis for model training or parameter calibration, thereby improving the accuracy of the simulation. After correction, the scenario simulation in step S2 is re-executed using the expanded historical traffic data, and subsequent steps are executed sequentially until the traffic congestion effect is satisfactory.

[0088] This invention addresses the issue of traffic diversion failure caused by prediction bias by re-evaluating the scenario simulation process when diversion fails after optimization during the decision-making process. It calculates the comprehensive deviation between the simulation and actual data and adjusts the sample size of historical traffic data based on this deviation in a positive correlation manner. The larger the comprehensive deviation, the greater the increase in sample size. This achieves closed-loop adaptive optimization of the scenario simulation process, enabling the simulation model to dynamically expand its training samples based on actual feedback, gradually improving simulation accuracy. This fundamentally solves the problem of diversion failure caused by prediction bias and enhances the overall robustness and long-term adaptability of highway traffic diversion methods during construction.

[0089] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for managing highway traffic during construction based on data prediction, characterized in that, include: Step S1: Obtain construction engineering data and preliminarily determine the scope of construction impact. Divide the scope of construction impact into several sub-regions according to the preset spatial grid size, and collect historical traffic data in each sub-region. Step S2: Based on the historical traffic data, a prediction method is used to extrapolate the scenario without traffic diversion, and the extrapolated scenario is obtained. Step S3: Based on the simulated scenario, construct a relocation decision for each sub-region and determine the expected data after implementing the relocation decision in each sub-region; Step S4: Begin construction and simultaneously implement the relocation decisions for each sub-area; Step S5: Periodically monitor the actual data of each sub-area within the construction impact range; Step S6: For each sub-region, calculate the matching degree of the sub-region based on the matching of the expected data and the actual data, and then calculate the comprehensive matching degree of all sub-regions. Based on the comprehensive matching degree, determine whether the traffic diversion during the current construction period is qualified. Step S7: For the current inspection cycle where traffic diversion is deemed unqualified during the current construction period, emergency diversion shall be implemented, and after the emergency diversion is completed, the scenario simulation process and / or diversion decision construction process shall be optimized.

2. The method for high-speed traffic diversion during construction based on data prediction according to claim 1, characterized in that, Calculating the overall matching degree of all sub-regions includes: For each sub-region, the matching degree of the sub-region is calculated based on the matching between the expected traffic density and the actual traffic density of the sub-region, as well as the matching between the expected vehicle type distribution and the actual vehicle type distribution of the sub-region. The overall matching score is obtained by weighted summing of the matching scores of all sub-regions. The weighting coefficient of each sub-region is determined based on the distance between the sub-region and the construction area, the historical average traffic flow density of the sub-region, the influencing factor, and the entanglement factor. The influencing factor represents the number of other sub-regions that can be affected after congestion occurs in the sub-region, and the entanglement factor represents the number of other sub-regions that can affect the traffic flow density of the sub-region.

3. The method for high-speed traffic diversion during construction based on data prediction according to claim 1, characterized in that, Determining whether traffic diversion is adequate during the current construction period based on the comprehensive matching degree includes: Compare the calculated overall matching score with the expected overall matching score; If the overall matching degree is greater than or equal to the expected overall matching degree, then the traffic diversion during the current construction period is deemed qualified. If the overall matching degree is less than the expected overall matching degree, the traffic diversion during the current construction period is deemed unqualified, emergency diversion is implemented, and the reason for the unqualified traffic diversion during the current construction period is determined after the emergency diversion is completed.

4. The method for high-speed traffic diversion during construction based on data prediction according to claim 3, characterized in that, The emergency evacuation operation includes: Sub-regions with a matching degree less than the expected matching degree are marked as sub-regions to be cleared; For each sub-region to be relocated, the relocation requirements of the sub-region to be relocated are determined based on the distance between the sub-region to be relocated and other sub-regions to be relocated, as well as the difference in matching degree between the sub-regions to be relocated; the relocation requirements are broken down into multiple sub-requirements; For each sub-area to be diverted, qualified surrounding sub-areas are matched according to its sub-demands, and traffic flow is diverted to different qualified surrounding sub-areas in batches.

5. The method for high-speed traffic diversion during construction based on data prediction according to claim 4, characterized in that, Before diverting traffic flow to different qualified surrounding sub-areas in batches, the following steps are also included: For each qualified sub-region with a matching degree greater than or equal to the expected matching degree, calculate the unpacking value of the qualified sub-region; The evacuation value is corrected based on the distance between the qualified sub-region and the sub-region to be evacuated; wherein, the greater the distance, the greater the correction magnitude of the evacuation value; If the corrected relief value is greater than or equal to the sub-demand of the sub-region to be relieved, then the qualified sub-region is determined to be able to relieve the sub-region to be relieved. If the corrected relocation value is less than the sub-requirement, then the qualified sub-region is determined not to be relocated from the sub-region to be relocated.

6. The method for high-speed traffic diversion during construction based on data prediction according to claim 4, characterized in that, After the emergency evacuation is completed, determine the reasons for the current construction period's traffic evacuation failures, including: For each sub-region to be relocated, obtain the distance between it and the nearest sub-region to be relocated, and record this distance as the shortest neighbor distance for that sub-region to be relocated; After obtaining the shortest neighbor distance for all sub-regions to be cleared, calculate the variance of each shortest neighbor distance; If the variance is less than the preset variance, the reason for the failure is determined to be in the scenario simulation process; If the variance is greater than or equal to the preset variance, the reason for the failure is determined to be in the process of constructing the decision-making process.

7. The method for high-speed traffic diversion during construction based on data prediction according to claim 6, characterized in that, When the reason for non-compliance lies in the decision-making process for relocation, the decision-making process for relocation should be optimized. Calculate the relocation decision application rate, which is the ratio of the number of sub-regions that have applied the relocation decision to the total number of sub-regions. The expected data for each sub-region is adjusted based on the application rate of the relocation decision; wherein, the higher the application rate of the relocation decision, the greater the increase in the expected data.

8. The method for high-speed traffic diversion during construction based on data prediction according to claim 7, characterized in that, Optimizing the decision-making process for relocation also includes: Calculate the overall relocation ratio; for each sub-area to be relocated, calculate the ratio of the traffic density eliminated after the emergency relocation decision to the difference between the actual density and the expected density, and record this ratio as the relocation ratio. The proportion of each relocation is corrected using a specific coefficient, and the average of the corrected proportions is recorded as the comprehensive relocation proportion; the specific coefficient is determined based on the relevant parameters of each sub-region to be relocated. The expected data is revised based on the overall relocation ratio; wherein, the larger the overall relocation ratio, the smaller the increase in the expected data.

9. The method for high-speed traffic diversion during construction based on data prediction according to claim 7, characterized in that, If the reason for non-compliance lies in the traffic diversion decision-making process and the traffic diversion decision-making process is optimized, and the traffic diversion is still deemed non-compliant during the current construction period, then the reason for non-compliance will be re-determined to lie in the scenario simulation process.

10. The method for high-speed traffic diversion during construction based on data prediction according to claim 6 or 9, characterized in that, When the reason for non-compliance lies in the scenario simulation process, the scenario simulation process should be optimized, including: The comprehensive deviation is calculated based on the simulated traffic flow density and vehicle type distribution in the simulation scenario, and the actual traffic flow density and vehicle type distribution obtained from actual detection. The sample size of the historical traffic data is adjusted based on the comprehensive deviation; wherein, the larger the comprehensive deviation, the greater the increase in the sample size of the historical traffic data.

Images