Expressway service area capacity dynamic regulation method for improving main line traffic efficiency

Abstract

This application discloses a method for dynamically controlling the capacity of highway service areas to improve mainline traffic efficiency. The method includes: acquiring mainline traffic operation data and service area operation data; determining whether to activate a control strategy based on a first preset condition; the first preset condition includes: activating the control strategy when the ratio of actual traffic flow to maximum road capacity exceeds a preset road traffic saturation threshold, and the average vehicle speed is lower than a preset speed threshold; if activated, implementing a vehicle guidance or closure decision for the service area based on its saturation, and issuing corresponding information prompts through an information dissemination device; the decision-making process includes: calculating the ratio of actual usage to maximum service capacity of the service areas in both directions, and implementing a vehicle guidance decision when at least one ratio is less than a preset saturation threshold; and implementing a closure decision when both ratios reach or exceed the preset saturation threshold. This application can improve mainline traffic efficiency and ensure safe operation.

Description

Technical Field

[0001] This application relates to the technical field of traffic operation control, specifically to a method for dynamic control of highway service area capacity to improve mainline traffic efficiency. Background Technology

[0002] Service areas, as crucial nodes on highways, not only provide static services such as vehicle parking, refueling, and catering, but their capacity utilization and opening / closing strategies also significantly impact traffic flow. However, current service area management and usage are relatively passive, lacking a dynamic control mechanism based on traffic congestion and service area saturation. Therefore, it is necessary to implement dynamic capacity control for highway service areas to reduce traffic volume on congested sections and improve mainline traffic efficiency.

[0003] In existing technologies, various methods have been employed to address highway traffic congestion and service area management issues. Traditionally, management relies on the experience of traffic police, who make judgments and decisions based on on-site conditions. Some patents focus on congestion prediction and resource allocation optimization within service areas. Examples include service area congestion prediction methods and information dissemination and guidance systems based on ETC data, which predict service area congestion through ETC data analysis and provide information guidance; and highway service area service resource allocation methods and systems based on vehicle-to-everything (V2X) technology, which utilize V2X technology to allocate service area resources. Other patents propose using artificial intelligence technology to predict vehicle information in the next moment to control service area opening and closing, employing reinforcement learning and other methods to formulate service area opening control decisions.

[0004] Clearly, existing highway service area operations and management generally suffer from passive and isolated management, unbalanced resource utilization, reliance on manual experience for management decisions, lack of automated triggering mechanisms, vague judgment criteria, and a lack of quantifiable collaborative judgment logic. There are no unified and executable rules for determining the level of congestion on the main line that necessitates intervention. Furthermore, existing technologies primarily focus on resolving congestion within the service area itself or on localized resource allocation from a single perspective, failing to consider collaborative strategies that combine main line traffic efficiency with service area saturation. This results in an inability to effectively intervene in main line traffic flow when facing complex traffic conditions, and a failure to significantly improve traffic efficiency in congested sections. Summary of the Invention

[0005] To address the issues of slow response and limited diversion effectiveness when relying solely on traffic police experience for traffic control in situations such as traffic accidents, severe weather, or serious congestion on existing highways, this application provides a dynamic capacity control method for highway service areas to improve mainline traffic efficiency and ensure the safe operation of highways.

[0006] Firstly, this application provides a method for dynamically adjusting the capacity of highway service areas to improve mainline traffic efficiency, including: Acquire traffic operation data for the main line; the traffic operation data includes: actual traffic flow, maximum road capacity, and average vehicle speed. Acquire service area operation data for the bidirectional service area; the service area operation data includes: actual usage and maximum service volume; Whether to activate the control strategy is determined based on the first preset condition; the first preset condition includes: when the ratio of actual traffic flow to maximum road capacity exceeds the preset road traffic saturation threshold and the average vehicle speed is lower than the preset speed threshold, the control strategy is activated. If enabled, the system will make decisions to guide or close the service area based on its saturation level and will issue corresponding information prompts through the information dissemination device. The decision-making based on the service area saturation level includes: calculating the ratio of the actual usage to the maximum service capacity of the two-way service area, and making a vehicle guidance decision when at least one ratio is less than a preset saturation threshold; and making a closure decision when both ratios reach or exceed the preset saturation threshold.

[0007] By adopting the above scheme, the mainline operation status and service area carrying capacity are jointly determined. Based on the mainline traffic operation data, the decision on whether to activate the control strategy is made in a precise quantitative manner. The saturation of the two-way service areas is used as the core safety constraint to avoid ineffective diversion or secondary congestion, and to dynamically balance the distribution of vehicles between the mainline and service areas.

[0008] Preferred, including: Continuously monitor the operation data of the main line and service areas to determine whether the second preset condition is met; the second preset condition includes: the ratio of main line traffic flow to the maximum road capacity is lower than the preset road traffic saturation threshold, the average vehicle speed is higher than the preset speed threshold, and the actual usage of both service areas exceeds the maximum service capacity; if the conditions are met, stop the vehicle guidance decision.

[0009] By adopting the above scheme, the operation data of the main line and service areas are continuously monitored. Based on the main line traffic status and service area usage, unnecessary vehicle guidance decisions are stopped in a timely manner. This ensures that the withdrawal of vehicle guidance decisions matches the recovery of main line traffic and the capacity of service areas, guaranteeing the timeliness and safety of the strategy withdrawal.

[0010] Preferred options also include: Acquire seasonal characteristics, holiday patterns, or road segment characteristics data, and dynamically adjust preset road traffic saturation thresholds and preset speed thresholds based on the acquired seasonal characteristics, holiday patterns, or road segment characteristics data.

[0011] By adopting the above scheme, the preset road traffic saturation threshold and preset speed threshold are adaptively optimized by combining seasonal characteristics, holiday travel patterns and the operating characteristics of specific road sections, thereby improving the adaptability and precision of the strategy.

[0012] Preferably, it further includes: acquiring and executing the optimal control variable in real time when making a vehicle guidance decision or a shutdown decision; the step of acquiring the optimal control variable in real time includes: Real-time acquisition of mainline congestion status data, two-way service area saturation status data, historical data of vehicles stuck in mainline congestion service areas, and correlation information between various service areas, as well as meteorological data; A hierarchical collaborative model architecture for the main line and service areas is constructed; the hierarchical collaborative model includes a prediction layer, a state evaluation layer, and an optimization decision layer; the optimal control variables are calculated in real time using the hierarchical collaborative model architecture for the main line and service areas; the calculation steps include: Based on real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, and correlation information between various service areas and meteorological data, the prediction layer is used to complete the prediction of main line congestion evolution, service area occupancy prediction, and real-time prediction of vehicle stay intention. Combining the prediction of mainline congestion evolution, service area occupancy, and real-time prediction of vehicle dwell intention, the state assessment layer is used to complete the assessment, and the assessment obtains the mainline traffic efficiency index, service area pressure index, and supply-demand matching index. A multi-objective optimization model is employed to solve for the optimal control variables. The solution process includes: defining control variables such as a guidance activation flag (indicating whether guidance should be activated for the i-th service area), a guidance rate (indicating the proportion of vehicles recommended to enter the i-th service area), and a closure flag (indicating whether the i-th service area should be closed); designing an objective function, which is obtained by weighted calculation based on maximizing mainline traffic efficiency, minimizing service area overflow risk, and minimizing interference with user journeys; defining constraints such as guidance rate constraint, service area capacity constraint, service area closure logic constraint, and time smoothing constraint; and solving for the optimal control variables at the current moment and converting them into control command outputs.

[0013] By adopting the above scheme, based on real-time acquired data, a hierarchical collaborative model architecture is used to accurately predict and evaluate indicators, and a multi-objective optimization model is used to solve for the optimal control variables and output control commands, thereby guiding vehicles more accurately and efficiently, achieving a better dynamic balance in vehicle allocation between the main line and service areas, and significantly improving the traffic efficiency of the main line.

[0014] Preferred options also include: After the monitoring and control commands are executed, the changes in the average speed of mainline vehicles and the service area occupancy rate are compared with the predicted values ​​to obtain deviation data. The deviation data is then input into the hierarchical collaborative model to retrain and adjust the parameters of each layer of the model.

[0015] By adopting the above scheme, the parameters of the hierarchical collaborative model are adjusted according to the deviation between the actual situation after the control command is executed and the predicted value, so as to make the model prediction more accurate and improve the accuracy and effectiveness of dynamic control of highway service area capacity.

[0016] Preferred, including: The congestion status data includes: congestion index, queue length, traffic density, congestion duration, main vehicle types and road segments experiencing congestion; the saturation status data includes: saturation area type, saturation area occupancy rate, and estimated waiting time in saturation areas; the related information includes: spatial distance and travel time between service areas, service area attribute complementarity data, and historical related traffic flow.

[0017] By adopting the above approach, more comprehensive and detailed data can be obtained, thereby more accurately assessing the mainline congestion status, service area saturation, and related information, and ultimately improving the efficiency of mainline traffic and the utilization efficiency of service area resources.

[0018] Preferred options also include: Considering the uncertainty of control variables, a predictive model design with coupled control variables is adopted, and a variational autoencoder is applied as the model architecture for the predictive layer. The model input includes: several candidate control variables as condition variables, and real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical dwell time data, correlation information between various service areas, and meteorological data. The model output includes: the main line congestion index for the preset future time period as the target variable, and the probability distribution of the occupancy rate of each service area in the future time period. The training process includes: extracting a large number of samples from historical control data as data preparation, each sample including: all condition variables at the corresponding time, and the target variable for the preset future time period starting from the corresponding time; initially setting the parameters of the encoder, decoder, and loss function in the model architecture; and training the predictive layer model structure using sample data. The current moment is used as a conditional variable. Several candidate control variables, along with real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, correlation information between various service areas, and meteorological data are input into the prediction layer to predict the main line congestion index, congestion probability, and occupancy probability distribution of each service area for the future period.

[0019] By adopting the above scheme, and considering the uncertainty of control variables, the variational autoencoder can accurately predict the probability distribution of the main line congestion index, congestion probability, and occupancy rate of each service area within a preset future time period, thereby enhancing the adaptability and accuracy of the control strategy.

[0020] Preferably, the mainline traffic efficiency indicators, service area pressure indicators, and supply-demand matching indicators are evaluated and obtained as follows: The mainline traffic efficiency index is defined as a comprehensive index of the quality of mainline traffic operation in the current and future periods, and the formula is: In the formula, This represents the main line traffic efficiency indicator. This indicates the current average speed of the main line. Indicates the current free-flow velocity of the main line; This represents the probability distribution of the average speed of the main line in the future time period obtained from the prediction layer, and the expected speed is calculated. This represents the probability distribution of the average speed of the main line in the future time period obtained from the prediction layer, and calculates the probability that the average speed is lower than a preset speed threshold. , as well as These are all weights of the corresponding parameters, and their sum is 1; The service area pressure index is defined as a comprehensive index of the current and future resource scarcity of the service area, and the formula is: In the formula, represents the pressure index of the i-th service area; k represents the functional area of ​​the i-th service area; This represents the occupancy rate of the k-th functional area in the i-th service area; This represents the expected future time-period occupancy rate of the k-th functional area of ​​the i-th service area obtained from the prediction layer; This represents the probability that the future time-period occupancy rate of the k-th functional area of ​​the i-th service area obtained from the prediction layer will exceed the preset occupancy rate threshold; , as well as All of these are the weights of the corresponding parameters, and their sum is 1; The weights of the pressure index corresponding to the k-th functional area of ​​the i-th service area are represented, and the sum is 1. The supply-demand matching index is defined as a measure of the balance between the demand of mainline vehicles for service area resources and the available capacity of service areas. The formula is: In the formula, D represents the total expected number of vehicles arriving from upstream that are willing to enter any service area, estimated using the stay intention prediction model. This indicates the total remaining capacity of all service areas. This also includes: calculating additional risk indicators, defined as the probability of adverse events occurring due to uncertainty, using the following formula: In the formula, This indicates the probability that the average speed of the main line obtained from the prediction layer is lower than the preset speed threshold for severe congestion; This represents the maximum probability of any service area or functional area overflowing during the prediction period.

[0021] By adopting the above scheme, the main line traffic efficiency, service area pressure, supply and demand matching degree and risks can be accurately quantified to complete a precise assessment, which can help to verify whether it is really necessary to start the control strategy.

[0022] Preferred options also include: A feasibility assessment is conducted to verify whether the control strategy truly needs to be activated; if the control strategy truly needs to be activated, a candidate set of service area control measures is determined; the feasibility assessment step includes: The system sets mainline congestion triggering conditions, service area capacity feasibility conditions, and supply-demand matching and risk conditions. The mainline congestion triggering conditions include: when the average speed of vehicles on the mainline is lower than a preset speed threshold and the duration exceeds a preset duration, or when the probability that the average speed of vehicles on the mainline will be lower than the preset speed threshold in a future period is greater than a preset probability threshold, and the predicted expected speed is lower than the preset speed threshold, then mainline congestion is considered to have been triggered; if the average speed of vehicles on the mainline is lower than the preset speed threshold and the duration exceeds a preset duration, and the probability that the average speed of vehicles on the mainline will be lower than the preset speed threshold in a preset future period is less than the preset probability threshold, then mainline congestion is considered not triggered. The service area capacity feasibility conditions include: if at least one service area pressure index value is less than the preset service area pressure index first threshold, and the preset multiple of the service area's predicted occupancy rate for the preset future period is less than 1, it is still lower than the preset occupancy rate threshold, then the service area is deemed to have guiding feasibility; if all service area pressure index values ​​are greater than the preset service area pressure index second threshold, and the probability that the occupancy rate of any functional area of ​​the service area exceeds the preset occupancy rate threshold for the preset future period is greater than the preset probability threshold, then the service area is deemed to be closed. The supply and demand matching and risk conditions include: if This indicates that demand exceeds supply, necessitating guidance for some vehicles to bypass all service areas; therefore, no adjustments are needed for any service areas. This indicates that congestion on the main line is worsening or there is a high risk of overflow from service areas; therefore, no adjustments are needed for any service areas. To preset risk thresholds; Based on all the conditions mentioned, determine whether there is a situation where the main line is congested, there are open service areas, and there are no service areas that do not require regulation at the current moment; if so, verify that regulation strategy really needs to be started, and determine the candidate set of service areas to participate in regulation based on the identified open service areas; otherwise, verify that regulation strategy does not need to be started and do not start regulation strategy.

[0023] By adopting the above scheme, a feasibility assessment is conducted to determine whether a control strategy truly needs to be implemented, thus avoiding unnecessary control measures. Conditions for triggering congestion on the main line, the feasibility of service area capacity, and supply-demand matching and risk conditions are set. By comprehensively considering factors such as main line traffic conditions, service area capacity, and supply-demand risks, a precise judgment is made on whether to implement a control strategy.

[0024] Preferred multi-objective optimization models include the following formulas for solving the optimal control variables: Assume there are currently N service areas participating in the regulation; the decision variables include the boot flags of each server. Guidance rate and the close sign ,in, Setting the value to 1 or 0 indicates whether to enable guidance for the i-th service area; Take a continuous variable [0,1] to represent whether the proportion of vehicles in the i-th service area is considered; Taking values ​​of 1 or 0 indicates whether to close the entrance to the i-th service area; the objective function is: In the formula, , , These are the weighting coefficients; In the formula, It is the average speed of the main line at the t-th time step in the future, where T is the prediction duration, and the current state refers to the traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, correlation information between various service areas, and meteorological data obtained at the current moment; ; In the formula, The sum of squares of the guiding laws; The constraints are: Guiding rate constraint: ; Service area capacity constraints: In the formula, The current flow rate is the upstream cross-section flow rate of the i service areas of the mainline. This represents the average probability that a vehicle will stay in the i-th service area in the predicted future time period. This represents the total number of services that can be accommodated by the i-th service area of ​​the main line and all downstream service areas at the current moment. Service area shutdown logic constraints: A service area Then its guiding rate must be ; Time smoothing constraint: In the formula, To limit the magnitude of changes in the preset guidance rate at adjacent decision-making moments; A heuristic algorithm is used to solve for the optimal control variables at the current moment and convert them into control command output.

[0025] By adopting the above scheme, taking into account the mainline traffic efficiency, service area spillover risk, and interference with users' journeys, the optimal control variables are solved using a multi-objective optimization model combined with constraints. This enables precise control of the guidance and closure decisions for multiple service areas, improving the mainline traffic efficiency and reducing the service area spillover risk, while also minimizing interference with users' journeys.

[0026] In summary, this application has the following beneficial effects: 1. Taking into account both mainline traffic operation data and service area operation data, control strategies are activated based on preset conditions to effectively intervene in mainline traffic flow and improve mainline traffic efficiency; the decision to activate control strategies is determined by quantitative preset conditions, and service area vehicle guidance or closure decisions are made based on service area saturation to avoid ineffective diversion or secondary congestion. 2. To achieve more refined vehicle guidance, a hierarchical collaborative model architecture is constructed. This architecture is used to complete multiple predictions and assessments, obtain mainline traffic efficiency indicators, service area pressure indicators, and supply-demand matching indicators. Then, a multi-objective optimization model is used to solve for the optimal control variables and output control commands. The optimal control variables can be accurately calculated based on real-time data to dynamically balance the allocation of vehicles on the mainline and in service areas, thereby improving the overall traffic efficiency and safety of the expressway. Attached Figure Description

[0027] Figure 1 This is a flowchart of the method for dynamic control of highway service area capacity to improve mainline traffic efficiency as described in a specific embodiment; Figure 2 This is a detailed flowchart illustrating the method for dynamic control of highway service area capacity to improve mainline traffic efficiency as described in a specific embodiment. Figure 3 This is a flowchart illustrating the method for obtaining control variables in the dynamic control method for improving the mainline traffic efficiency of highway service areas, as described in a specific embodiment. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0029] like Figure 1 As shown in the figure, this application discloses a method for dynamic capacity control of highway service areas to improve mainline traffic efficiency, including: S1. Obtain traffic operation data for the main line of the expressway.

[0030] Specifically, the traffic operation data for the main line includes: actual traffic flow, maximum road capacity, and average vehicle speed. IoT devices such as microwave detectors and video detectors are deployed at key sections of the main line and ramp entrances / exits to collect real-time data on actual traffic flow and average vehicle speed across the main line cross-section. This data is then combined with AI algorithms to identify and obtain operational data such as vehicle type, queue length, and driving status in real time.

[0031] For example, microwave detectors use the principle of microwave reflection to detect vehicle speed and flow, while video detectors monitor the road through cameras, use image recognition technology to analyze vehicle movement, and collect real-time data on the actual traffic flow V and average vehicle speed at the main road section. And based on the road design standards, the maximum traffic capacity C of this main road section is determined.

[0032] S2. Obtain service area operation data for the two-way service area.

[0033] Specifically, the service area operation data includes actual usage and maximum service capacity. Service area status acquisition equipment may include parking space detectors, video surveillance, and other devices to obtain real-time actual usage and maximum service capacity of the two-way service area. Parking space detectors may employ ultrasonic sensors, geomagnetic sensors, etc., to determine the actual usage of the service area by detecting the occupancy of parking spaces; while video surveillance can monitor the overall situation of the service area in real time. Additionally, ETC data can be used to obtain vehicle entry and exit information for the service area.

[0034] For example, by using parking space detectors, video surveillance, and other equipment within the service area, the actual number of vehicles occupied in both directions of the service area can be obtained in real time. , and its maximum design capacity , .

[0035] S3. Determine whether to activate the control strategy based on the first preset condition.

[0036] Specifically, the first preset condition includes: when the mainline traffic status simultaneously meets the following two conditions, the dynamic control strategy for highway service areas will be activated, including: and

[0037] In the formula, M is the road traffic saturation threshold, usually taken as 0.8; K is the speed threshold, generally taken as 50 km / h. The road traffic saturation threshold M and speed threshold K can be adaptively optimized by combining seasonal characteristics, holiday travel patterns, and the operational characteristics of specific road sections, thereby improving the adaptability and precision of the strategy. For example, during holidays, due to the large traffic volume, the preset road traffic saturation threshold and preset speed threshold can be appropriately reduced to activate the control strategy earlier; in certain special road sections, such as mountainous sections or accident-prone sections, the thresholds can also be adjusted according to the actual situation.

[0038] If the ratio of the actual traffic flow to the maximum road capacity collected at the current moment exceeds the preset road traffic saturation threshold, and the average vehicle speed is lower than the preset speed threshold, then the current mainline traffic is considered congested, and the control strategy is activated.

[0039] S4. Based on the result of the judgment to activate the control strategy, make a decision to guide or close the service area based on the service area saturation, and issue corresponding information prompts through the information release device.

[0040] Specifically, such as Figure 2 As shown, the decision to guide vehicles to or close a service area is made based on its status. The method for making this decision is as follows: or

[0041] in, , The actual saturation of the two-way service area is represented by the following formula: ,

[0042] If the following conditions are met or If at least one service area has the capacity to receive vehicles, a vehicle guidance decision will be made; otherwise... or If both service areas are saturated, then a service area closure decision will be made.

[0043] Based on the above assessment, if the decision to guide vehicles to the service area is made, then vehicle guidance will be activated, and announcements such as "Congestion ahead, please enter the service area for rest" will be made through information boards, audio-visual equipment, and navigation software. If the decision to close the service area is made, then the service area will be closed, and announcements such as "Service area temporarily closed" will be gradually made through information boards, audio-visual equipment, and navigation software. Vehicles will no longer be guided to the opposite service area, and route avoidance prompts will be provided in the navigation system to prevent vehicles from continuing to enter the already saturated service area. The audio-visual equipment mentioned includes, but is not limited to, voice broadcasting equipment, high-brightness LED warning lights, flashing warning lights, variable message signs, and electronic screens in the service area. These devices work together to form a multi-layered, three-dimensional information dissemination and guidance network.

[0044] Furthermore, when determining the conditions for policy closure, both the recovery status of mainline traffic and the carrying capacity limits of the two-way service areas were considered, ensuring the timeliness and safety of policy withdrawal. Figure 2 As shown, the method further includes: S5. Continuously monitor the operation data of the main line and service areas to determine whether the second preset condition is met, and shut down the vehicle guidance decision and restore the normal operation of the service area when the second preset condition is met.

[0045] Specifically, during the implementation of the strategy, the system continuously monitors the operational status of the highway mainline and service areas, that is, it continuously acquires data such as the ratio of mainline traffic flow to the maximum road capacity, average vehicle speed, and actual usage of service areas in both directions.

[0046] The second preset condition includes: and and or If the ratio of mainline traffic flow to maximum road capacity is lower than the preset road traffic saturation threshold, the average vehicle speed is higher than the preset speed threshold, and the actual usage of both service areas exceeds the maximum service capacity, then the mainline traffic congestion has been alleviated and the actual usage of both service areas has exceeded the maximum service capacity. In this case, the vehicle guidance decision for the service area will be turned off. Otherwise, the vehicle guidance decision will not be turned off and the judgment will continue.

[0047] A specific embodiment, differing from the above embodiments, considers implementing vehicle guidance or closure decisions for service areas based on service area saturation, and calculates and obtains appropriate guidance and closure decision control variables. This better achieves a dynamic balance in vehicle allocation between the mainline and service areas, significantly improving mainline traffic efficiency. The method further includes: When executing vehicle guidance or shutdown decisions, the optimal control variables are acquired and executed in real time; the steps for acquiring the optimal control variables in real time include: S41. Obtain relevant data on decision control variables in real time.

[0048] Specifically, real-time congestion status data for the main line is acquired, including: continuous values ​​representing congestion levels such as congestion index (average speed and free flow), queue length, and traffic density (veh / km / lane); congestion duration (the current duration of congestion and the remaining duration predicted based on historical patterns); the main vehicle types involved in congestion (the proportion of trucks, buses, cars, and trolleybuses in the congested flow is determined through ETC gantry data or video recognition); and road segments. Specifically, this congestion status data can be collected simultaneously with the aforementioned main line traffic operation data collection and statistically calculated and identified at the initial congestion assessment point, thereby obtaining the main line's congestion status data.

[0049] Real-time saturation status data of the two-way service area is acquired, including: saturation area type, saturation area occupancy rate, and estimated waiting time in the saturation area. Each area refers to the service area being divided into parking areas (cars / trucks / buses), refueling areas, charging areas, and catering / rest areas. Combining video device data with the service area's back-end management and interactive system, the occupancy rate of each area is monitored in real time, and the queuing time in each area is estimated based on historical average service times.

[0050] Real-time acquisition of historical data on vehicle dwell times at service areas along the main road congestion route (e.g., vehicle type, trip duration, current time, holiday status, weather, service area brand / facility rating, etc.), correlation information between service areas, and meteorological data. The correlation information includes: spatial distance and travel time between service areas (mileage and estimated travel time between upstream and downstream service areas), service area attribute complementarity data (upstream service area charging saturation, downstream service area charging availability), and historical correlation traffic flow (probability of vehicle movement between two service areas in historical data).

[0051] S42. Construct a hierarchical collaborative model architecture for the main line and service areas.

[0052] Specifically, the hierarchical collaborative model includes a prediction layer, a state evaluation layer, and an optimization decision layer.

[0053] Firstly, the prediction layer is mainly used to complete the prediction of mainline congestion evolution (in order to obtain the mainline congestion coefficient, etc. in future time steps), service area occupancy prediction (in order to obtain the occupancy rate in future time steps), and real-time estimation of vehicle dwelling intentions (in order to obtain the distribution of group dwelling demand in future time steps).

[0054] Therefore, for mainline congestion prediction, graph neural networks or Transformer models can be used, combined with real-time traffic flow and event information, to predict the mainline congestion index for multiple future time steps. For service area occupancy prediction, based on the current occupancy rate of each area, historical arrival rate, and vehicle dwell intention prediction, time series models (such as Prophet and LSTM) are used to predict the future occupancy rate and queue length of each area. For real-time estimation of vehicle dwell intention, for vehicles on the current congested road segment, real-time vehicle characteristics (which can be identified through ETC gantries, such as vehicle type and new energy vehicle license plates) are input into the dwell probability model to obtain the distribution of group dwell demand. For example, by combining the historical dwell data of mainline congested vehicles at service areas, LSTM or other deep learning models can be used to obtain the dwell probability of vehicles coming from upstream of the congested road segment at each service area.

[0055] However, the above model predictions are usually based only on historical data and the current state, but ignore the feedback effect of the control variables to be executed (guidance, shutdown) on the future state, and the uncertainty brought about by this feedback. In order to obtain more accurate prediction results, it is necessary to construct a probabilistic prediction model that can couple control decisions with traffic evolution, such as: adopting a prediction model design that couples control variables and applying a variational autoencoder as the prediction layer model architecture.

[0056] The model inputs include: current mainline traffic status (real-time traffic operation data and congestion status data), current service area status (real-time service area operation data and service area saturation status data), other factors (interrelationship information between service areas and meteorological data), and control variables to be evaluated (candidate guidance rates and closure flags) as condition variables. The model output is: multiple future time steps. The target variables include the probability distribution of the main line congestion index (e.g., 5-minute, 15-minute, 30-minute), and the probability distribution of service area occupancy rates at multiple future time steps. The model is trained using historical control data, with historically occurring control variables as conditions, to train a conditional generative model capable of generating samples consistent with the true future state distribution. A variational autoencoder is used to learn the conditional probability distribution of the future state given the current state and control variables. Each sample includes: the corresponding time (e.g., ...). Under all conditional variables, the corresponding time period is preset for the future. The target variable; the encoder in the initial model architecture (variational autoencoder) is defined. decoder and the parameters of the loss function; where, Representing the future true state, condition c includes the current state and control variables, and outputs latent variables. The mean and variance, prior, Assuming a standard normal distribution; Loss function: Maximizing the lower bound of evidence for the conditional log-likelihood: In the formula, the first term is the reconstruction loss, and the second term is the KL divergence. The weights are used to train the prediction layer model structure using sample data.

[0057] Secondly, the state assessment layer is mainly used to quantify the current and predicted states as a basis for decision-making and to design indicators. Specific indicators include: mainline traffic efficiency indicators (current mainline average speed / free flow speed), service area pressure indicators, and supply and demand matching degree.

[0058] Third, the decision-making layer is optimized, primarily using a multi-objective optimization model to solve for the optimal control variables and transform them into control command outputs. The control variables are defined as follows: a guidance activation flag indicating whether guidance should be activated for the i-th service area; a guidance rate indicating the proportion of vehicles recommended to enter the i-th service area; and a closure flag indicating whether the i-th service area should be closed. The objective function is designed by weighting the calculation based on maximizing mainline traffic efficiency, minimizing service area overflow risk, and minimizing interference with user journeys. The constraints are defined as follows: guidance rate constraint, service area capacity constraint, service area closure logic constraint, and time smoothing constraint.

[0059] S43. Utilizing the hierarchical collaborative model architecture of the mainline and service areas, the optimal control variables are calculated and obtained in real time, and then converted into control commands for execution. The calculation steps include: First, based on real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, and correlation information between various service areas, as well as meteorological data, the prediction layer is used to complete the prediction of main line congestion evolution, service area occupancy prediction, and real-time prediction of vehicle stay intention.

[0060] Specifically, the current moment is used as a conditional variable for several candidate control variables, along with real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, correlation information between various service areas, and meteorological data, all of which are input into the prediction layer to predict the main line congestion index, congestion probability, and occupancy probability distribution of each service area for the future period.

[0061] Then, combining the mainline congestion evolution prediction, service area occupancy prediction, and real-time vehicle dwell intention prediction, the state assessment layer is used to complete the assessment, obtaining the mainline traffic efficiency index, service area pressure index, and supply-demand matching index; the specific formulas are as follows: The mainline traffic efficiency index is defined as a comprehensive index of the quality of mainline traffic operation in the current and future periods, and the formula is: In the formula, This represents the main line traffic efficiency indicator. This indicates the current average speed of the main line. Indicates the current free-flow velocity of the main line; This represents the probability distribution of the average speed of the main line in the future time period obtained from the prediction layer, and the expected speed is calculated. This represents the probability distribution of the average speed of the main line in the future time period obtained from the prediction layer, and calculates the probability that the average speed is lower than a preset speed threshold. , as well as All of these are the weights of the corresponding parameters, and their sum is 1; The service area pressure index is defined as a comprehensive index of the current and future resource scarcity of the service area, and the formula is: In the formula, represents the pressure index of the i-th service area; k represents the functional area of ​​the i-th service area; This represents the occupancy rate of the k-th functional area in the i-th service area; This represents the expected future time-period occupancy rate of the k-th functional area of ​​the i-th service area obtained from the prediction layer; This represents the probability that the future time-period occupancy rate of the k-th functional area of ​​the i-th service area obtained from the prediction layer will exceed the preset occupancy rate threshold; , as well as All of these are the weights of the corresponding parameters, and their sum is 1; The weights of the pressure index corresponding to the k-th functional area of ​​the i-th service area are represented, and the sum is 1. The supply-demand matching index is defined as a measure of the balance between the demand of mainline vehicles for service area resources and the available capacity of service areas. The formula is: In the formula, D represents the total expected number of vehicles arriving from upstream that are willing to enter any service area, estimated using the stay intention prediction model. This indicates the total remaining capacity of all service areas; it also includes: additionally calculated risk indicators, defined as the probability of adverse events occurring due to uncertainty, with the formula: In the formula, This indicates the probability that the average speed of the main line obtained from the prediction layer is lower than the preset speed threshold for severe congestion; This represents the maximum probability of any service area or functional area overflowing during the prediction period.

[0062] Then, before solving for the final optimal control variables, the feasibility assessment of the regulation is restarted to verify whether the regulation strategy truly needs to be activated. If the regulation strategy truly needs to be activated, the service area regulation candidate set is determined. The feasibility assessment verification steps include: The system sets the mainline congestion triggering conditions, service area capacity feasibility conditions, and supply-demand matching and risk conditions. Based on all these conditions, it determines whether there is a situation where the mainline is congested, there are open service areas, and there are no situations where all service areas do not require regulation. If such a situation exists, it verifies that a regulation strategy truly needs to be activated, and determines the candidate set of service areas to participate in regulation based on the identified open service areas. Otherwise, it verifies that no regulation strategy needs to be activated and does not activate the regulation strategy.

[0063] The mainline congestion triggering conditions include: when the average speed of vehicles on the mainline is lower than a preset speed threshold (60 km / h) and the duration exceeds a preset duration (e.g., 5 minutes), or when the probability that the average speed of vehicles on the mainline will be lower than the preset speed threshold in the future period is greater than a preset probability threshold (0.6), and the predicted expected speed is lower than the preset speed threshold, it indicates that congestion has occurred or is about to occur, and mainline congestion is determined to be triggered; if the average speed of vehicles on the mainline is lower than the preset speed threshold and the duration exceeds the preset duration, and the probability that the average speed of vehicles on the mainline will be lower than the preset speed threshold in the preset future period is less than the preset probability threshold, it is determined that mainline congestion has not been triggered, that is, it is currently congested but is predicted to ease soon, so no control is implemented for the time being.

[0064] The service area capacity feasibility conditions include: at least one service area pressure index value is less than the preset service area pressure index first threshold (0.8), and the preset multiple of the service area's predicted occupancy rate for the preset future period (90%) is still lower than the preset occupancy rate threshold, then the service area is deemed to have guiding feasibility; if all service area pressure index values ​​are greater than the preset service area pressure index second threshold (0.95), and the probability that the occupancy rate of any functional area of ​​the service area exceeds the preset occupancy rate threshold for the preset future period is greater than the preset probability threshold, then the service area is deemed to be closed to prevent the queue from spreading from the service area to the main line.

[0065] The supply and demand matching and risk conditions include: if This indicates that demand exceeds supply, and some vehicles must be guided to skip all service areas; no adjustments are needed for any service areas (i.e., the guidance rate is negative, indicating that the service areas are full). This indicates that congestion on the main line is worsening or there is a high risk of overflow from service areas; therefore, no adjustments are needed for any service areas. This is a preset risk threshold.

[0066] Finally, a multi-objective optimization model is used to solve for the optimal control variables. The solution process includes: Assume there are currently N service areas participating in the regulation; the decision variables include the boot flags of each server. Guidance rate and the close sign ,in, Setting the value to 1 or 0 indicates whether to enable guidance for the i-th service area; Take a continuous variable [0,1] to represent whether the proportion of vehicles in the i-th service area is considered; Taking values ​​of 1 or 0 indicates whether to close the entrance to the i-th service area; the objective function is: In the formula, , , These are the weighting coefficients; In the formula, It is the average speed of the main line at the t-th time step in the future, where T is the prediction duration, and the current state refers to the traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, correlation information between various service areas, and meteorological data obtained at the current moment; ; In the formula, The sum of squares of the guiding laws; The constraints are: Guiding rate constraint: ; Service area capacity constraints: In the formula, The current flow rate is the upstream cross-section flow rate of the i service areas of the mainline. This represents the average probability that a vehicle will stay in the i-th service area in the predicted future time period. This represents the total number of services that can be accommodated by the i-th service area of ​​the main line and all downstream service areas at the current moment. Service area shutdown logic constraints: A service area Then its guiding rate must be ; Time smoothing constraint: In the formula, To limit the magnitude of changes in the preset guidance rate at adjacent decision-making moments; A heuristic algorithm is used to solve for the optimal control variables at the current time step and convert them into control command outputs. Taking a genetic algorithm as an example, the decision variables are encoded into chromosomes. Fitness is calculated based on the objective function and constraint penalty terms, and evolutionary operations—selection, crossover, and mutation—are performed for iterative optimization.

[0067] S44. Conduct follow-up execution feedback and adjustments.

[0068] Specifically, after the monitoring and control commands are executed, the status of the main line (main line speed) and the changes in service area occupancy are collected and compared with the predicted values ​​to obtain deviation data, such as the actual average speed of the main line and the predicted average speed, and the actual service area occupancy rate and the predicted service area occupancy rate.

[0069] The system collects and feeds actual response data back to the prediction layer, using this data to update the model parameters. Based on the collected data from the actual execution cycle (upstream vehicle traffic, actual guidance rate, and actual vehicle traffic entering each service area), the actual response rate (actual response data based on guidance to service areas) is calculated. This actual response rate, along with its corresponding upstream vehicle traffic, actual guidance rate, and actual vehicle traffic entering each service area, is fed back to the original prediction layer as training data to train and update the parameters of the real-time estimation model for vehicle dwell intention.

[0070] Simultaneously, deviation data is input into the hierarchical collaborative model to adjust the parameters of each layer. For example, deviations between the predicted average speed and the actual average speed, or between the predicted occupancy rate and the actual occupancy rate, are fed back to the prediction layer as feedback data to train and adjust the parameters of the variational autoencoder, or to the decision layer as feedback data to train and adjust the weight parameters of the multi-objective optimization model. Furthermore, comprehensive indicators (mainline efficiency, service area pressure, supply-demand matching degree, and risk indicators) are generated based on predicted and real-time data, and their corresponding deviations are fed back to the state assessment layer to adjust the weights of the various indicators calculated in the state assessment layer.

[0071] This application also discloses a computer-readable storage medium.

[0072] Specifically, the computer-readable storage medium stores a computer program that can be loaded by a processor and executed, such as the aforementioned method for dynamically adjusting the capacity of highway service areas to improve mainline traffic efficiency. The computer-readable storage medium includes, for example, various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0073] This application also discloses a computer device.

[0074] Specifically, the computer device includes a memory and a processor. The memory stores a computer program that can be loaded by the processor and executed to improve the dynamic control method of highway service area capacity to enhance mainline traffic efficiency.

[0075] The above are all preferred embodiments of this application and are not intended to limit the scope of protection of this application. Any feature disclosed in this specification (including the abstract and drawings) may be replaced by other equivalent or similar features unless specifically stated otherwise. That is, unless specifically stated otherwise, each feature is only one example of a series of equivalent or similar features.

Claims

1. A method for dynamically adjusting the capacity of highway service areas to improve mainline traffic efficiency, characterized in that, include: Obtain traffic operation data for the main line; The traffic operation data includes: actual traffic flow, maximum road capacity, and average vehicle speed; Acquire service area operation data for the bidirectional service area; the service area operation data includes: actual usage and maximum service volume; Whether to activate the control strategy is determined based on the first preset condition; the first preset condition includes: when the ratio of actual traffic flow to maximum road capacity exceeds the preset road traffic saturation threshold and the average vehicle speed is lower than the preset speed threshold, the control strategy is activated. If enabled, the system will make decisions to guide or close the service area based on its saturation level and will issue corresponding information prompts through the information dissemination device. The decision-making based on the service area saturation level includes: calculating the ratio of the actual usage to the maximum service capacity of the two-way service area, and making a vehicle guidance decision when at least one ratio is less than a preset saturation threshold; and making a closure decision when both ratios reach or exceed the preset saturation threshold.

2. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 1, characterized in that, include: Continuously monitor the operational data of the main line and service areas to determine whether the second preset condition is met; The second preset condition includes: the ratio of mainline traffic flow to the maximum road capacity is lower than the preset road traffic saturation threshold, the average vehicle speed is higher than the preset speed threshold, and the actual usage of both directions of the service area exceeds the maximum service capacity; if these conditions are met, the vehicle guidance decision will be stopped.

3. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 2, characterized in that, Also includes: Acquire seasonal characteristics, holiday patterns, or road segment characteristics data, and dynamically adjust preset road traffic saturation thresholds and preset speed thresholds based on the acquired seasonal characteristics, holiday patterns, or road segment characteristics data.

4. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 2, characterized in that, Also includes: When making vehicle guidance or shutdown decisions, the optimal control variables are acquired and executed in real time; The steps for real-time acquisition of the optimal control variable include: Real-time acquisition of mainline congestion status data, two-way service area saturation status data, historical data of vehicles stuck in mainline congestion service areas, and correlation information between various service areas, as well as meteorological data; A hierarchical collaborative model architecture for the main line and service areas is constructed; the hierarchical collaborative model includes a prediction layer, a state evaluation layer, and an optimization decision layer; the optimal control variables are calculated in real time using the hierarchical collaborative model architecture for the main line and service areas; the calculation steps include: Based on real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, and correlation information between various service areas and meteorological data, the prediction layer is used to complete the prediction of main line congestion evolution, service area occupancy prediction, and real-time prediction of vehicle stay intention. Combining the prediction of mainline congestion evolution, service area occupancy, and real-time prediction of vehicle dwell intention, the state assessment layer is used to complete the assessment, and the assessment obtains the mainline traffic efficiency index, service area pressure index, and supply-demand matching index. A multi-objective optimization model is employed to solve for the optimal control variables. The solution process includes: defining control variables such as a guidance activation flag (indicating whether guidance should be activated for the i-th service area), a guidance rate (indicating the proportion of vehicles recommended to enter the i-th service area), and a closure flag (indicating whether the i-th service area should be closed); designing an objective function, which is obtained by weighted calculation based on maximizing mainline traffic efficiency, minimizing service area overflow risk, and minimizing interference with user journeys; defining constraints such as guidance rate constraint, service area capacity constraint, service area closure logic constraint, and time smoothing constraint; and solving for the optimal control variables at the current moment and converting them into control command outputs.

5. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 4, characterized in that, Also includes: After the monitoring and control commands are executed, the changes in the average speed of mainline vehicles and the service area occupancy rate are compared with the predicted values ​​to obtain deviation data. The deviation data is then input into the hierarchical collaborative model to retrain and adjust the parameters of each layer of the model.

6. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 4, characterized in that, include: The congestion status data includes: congestion index, queue length, congestion duration, main vehicle types and road segments involved in congestion; the saturation status data includes: saturation area type, saturation area occupancy rate, and estimated waiting time in saturation areas; the related information includes: spatial distance and travel time between service areas, complementary attribute data of service areas, and historical related traffic.

7. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 6, characterized in that, Also includes: Consider the uncertainty of the control variables; A predictive model with coupled control variables is designed, and a variational autoencoder is applied as the model architecture for the predictive layer. The model input includes: several candidate control variables as condition variables, and real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical dwell time data, correlation information between various service areas, and meteorological data; the model output includes: the main line congestion index for the preset future time period as the target variable, and the probability distribution of the occupancy rate of each service area in the future time period; the training process includes: extracting a large number of samples from historical control data for data preparation, each sample including: all condition variables at the corresponding time, and the target variable for the preset future time period starting from the corresponding time; initially setting the parameters of the encoder, decoder, and loss function in the model architecture; and training the prediction layer model structure using the sample data; The current moment is used as a conditional variable. Several candidate control variables, along with real-time traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, correlation information between various service areas, and meteorological data are input into the prediction layer to predict the main line congestion index, congestion probability, and occupancy probability distribution of each service area for the future period.

8. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 7, characterized in that, The assessment includes obtaining indicators such as mainline traffic efficiency, service area pressure, and supply-demand matching. The mainline traffic efficiency index is defined as a comprehensive index of the quality of mainline traffic operation in the current and future periods, and the formula is: In the formula, This represents the main line traffic efficiency indicator. This indicates the current average speed of the main line. Indicates the current free-flow velocity of the main line; This represents the probability distribution of the average speed of the main line in the future time period obtained from the prediction layer, and the expected speed is calculated. This represents the probability distribution of the average speed of the main line in the future time period obtained from the prediction layer, and calculates the probability that the average speed is lower than a preset speed threshold. , as well as These are all weights of the corresponding parameters, and their sum is 1; The service area pressure index is defined as a comprehensive index of the current and future resource scarcity of the service area, and the formula is: In the formula, represents the pressure index of the i-th service area; k represents the functional area of ​​the i-th service area; This represents the occupancy rate of the k-th functional area in the i-th service area; This represents the expected future time-period occupancy rate of the k-th functional area of ​​the i-th service area obtained from the prediction layer; This represents the probability that the future time-period occupancy rate of the k-th functional area of ​​the i-th service area obtained from the prediction layer will exceed the preset occupancy rate threshold; , as well as These are all weights of the corresponding parameters, and their sum is 1; The weights of the pressure index corresponding to the k-th functional area of ​​the i-th service area are represented, and the sum is 1. The supply-demand matching index is defined as a measure of the balance between the demand of mainline vehicles for service area resources and the available capacity of service areas. The formula is: In the formula, D represents the total expected number of vehicles arriving from upstream that are willing to enter any service area, estimated using the stay intention prediction model. This indicates the total remaining capacity of all service areas. This also includes: calculating additional risk indicators, defined as the probability of adverse events occurring due to uncertainty, using the following formula: In the formula, This indicates the probability that the average speed of the main line obtained from the prediction layer is lower than the preset speed threshold for severe congestion; This represents the maximum probability of any service area or functional area overflowing during the prediction period.

9. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 8, characterized in that, Also includes: Conduct a feasibility assessment to verify whether it is truly necessary to initiate regulatory measures. If it is truly necessary to activate the control strategy, then determine the service area control candidate set; The feasibility assessment and verification steps include: The system sets mainline congestion triggering conditions, service area capacity feasibility conditions, and supply-demand matching and risk conditions. The mainline congestion triggering conditions include: when the average speed of vehicles on the mainline is lower than a preset speed threshold and the duration exceeds a preset duration, or when the probability that the average speed of vehicles on the mainline will be lower than the preset speed threshold in a future period is greater than a preset probability threshold, and the predicted expected speed is lower than the preset speed threshold, then mainline congestion is considered to have been triggered; if the average speed of vehicles on the mainline is lower than the preset speed threshold and the duration exceeds a preset duration, and the probability that the average speed of vehicles on the mainline will be lower than the preset speed threshold in a preset future period is less than the preset probability threshold, then mainline congestion is considered not triggered. The service area capacity feasibility conditions include: if at least one service area pressure index value is less than the preset service area pressure index first threshold, and the preset multiple of the service area's predicted occupancy rate for the preset future period is less than 1, it is still lower than the preset occupancy rate threshold, then the service area is deemed to have guiding feasibility; if all service area pressure index values ​​are greater than the preset service area pressure index second threshold, and the probability that the occupancy rate of any functional area of ​​the service area exceeds the preset occupancy rate threshold for the preset future period is greater than the preset probability threshold, then the service area is deemed to be closed. The supply and demand matching and risk conditions include: if This indicates that demand exceeds supply, necessitating guidance for some vehicles to bypass all service areas; therefore, no adjustments are needed for any service areas. This indicates that congestion on the main line is worsening or there is a high risk of overflow from service areas; therefore, no adjustments are needed for any service areas. To preset risk thresholds; Based on all the conditions mentioned, determine whether there is a situation where the main line is congested, there are open service areas, and there are no service areas that do not require regulation at the current moment; if so, verify that regulation strategy really needs to be started, and determine the candidate set of service areas to participate in regulation based on the identified open service areas; otherwise, verify that regulation strategy does not need to be started and do not start regulation strategy.

10. The method for dynamic control of highway service area capacity to improve mainline traffic efficiency according to claim 9, characterized in that, For multi-objective optimization models, the formulas for solving for the optimal control variables include: Assume there are currently N service areas participating in the regulation; the decision variables include the boot flags of each server. Guidance rate and the close sign ,in, Setting the value to 1 or 0 indicates whether to enable guidance for the i-th service area; Take a continuous variable [0,1] to represent the proportion of vehicles in the i-th service area; Taking values ​​of 1 or 0 indicates whether to close the entrance to the i-th service area; the objective function is: In the formula, , , These are the weighting coefficients; In the formula, It is the average speed of the main line at the t-th time step in the future, where T is the prediction duration, and the current state refers to the traffic operation data, service area operation data, congestion status operation data, saturation status data, historical stay data, correlation information between various service areas, and meteorological data obtained at the current moment; ； In the formula, The sum of squares of the guiding laws; The constraints are: Guiding rate constraint: ; Service area capacity constraints: In the formula, The current flow rate is the upstream cross-section flow rate of the i service areas of the mainline. This represents the average probability that a vehicle will stay in the i-th service area in the predicted future time period. This represents the total number of services that can be accommodated by the i-th service area of ​​the main line and all downstream service areas at the current moment. Service area shutdown logic constraints: A service area Then its guiding rate must be ; Time smoothing constraint: In the formula, To limit the magnitude of the preset guidance rate change at adjacent decision-making moments; A heuristic algorithm is used to solve for the optimal control variables at the current moment and convert them into control command output.

Images