Dynamic mobile overload vehicle patrol path planning method and system based on demand prediction

Abstract

This invention provides a dynamic mobile overload control vehicle route planning method and system based on demand forecasting, comprising: first, processing historical overload control data; second, constructing an overload control demand forecasting model, inputting a truck spatiotemporal feature matrix containing time and spatial information into the overload control demand forecasting model to predict truck traffic in the next time slice, thereby determining the change of overload control demand points; third, inputting the overload control demand points into an overload control vehicle patrol route planning model with the objective of minimizing the patrol distance cost and time window penalty cost of overload control vehicles in the next time slice, to obtain the dynamic overload control vehicle patrol route within the overload control area. This method improves upon the current reliance on fixed overload control stations for overload and oversize vehicle management, effectively preventing overloaded trucks from detouring to avoid overload control stations, improving overload control efficiency and effectiveness while reducing costs and mitigating the hazards of overload and oversize vehicles.

Description

Technical Field

[0001] This invention relates to the field of highway network overload control technology, and in particular to a dynamic mobile overload control vehicle route planning method and system based on demand forecasting.

[0002] Overloaded transport refers to the act of vehicles exceeding the limits for axle load or total weight on highways or highway structures. Overloaded transport causes significant damage to highway infrastructure. Currently, the management of overloaded trucks on highways mainly relies on fixed checkpoints, but this fixed approach is often insufficient to effectively intercept overloaded trucks. Because highway networks are complex and the number of checkpoints is limited, mobile enforcement vehicles are needed to control and manage these overloaded trucks. Research on overload control and vehicle route planning mainly includes the following:

[0003] This study analyzes the spatial distribution of overload control stations within the road network. Besinovic et al., aiming to minimize costs (including investment and construction costs of the stations and expected damage costs to infrastructure from overloaded trucks), established an optimization model for the number and location of overload control stations based on stochastic traffic flow. E. MORENO-QUINTERO et al. developed a bilevel programming model with the upper level aiming to minimize the operating costs of administrators and the lower level aiming to minimize the transportation costs of transport operators. This model describes the response of transport operators to changes in enforcement density and fines, identifying the overloaded transport situation and the placement of overload control stations under equilibrium conditions.

[0004] Vehicle routing research primarily focuses on delivery problems. Hannan MA et al. studied vehicle routing problems with capacity constraints; Min et al. studied multi-vehicle routing optimization problems involving simultaneous pickup and delivery. Ghannadpour et al. studied a multi-objective dynamic vehicle routing problem with fuzzy time windows. Schyns considered changes in customer demand, time windows, and customer locations, aiming to minimize vehicle response time, and studied a multi-vehicle dynamic vehicle routing problem with capacity and time window constraints.

[0005] Based on a review of relevant domestic and international literature, the limitations of existing research on highway overload control are summarized as follows: In the study of overload and oversize control in highway networks, most studies only focus on the deployment of fixed overload control stations, without considering the advantages of mobile overload control modes, making it difficult to effectively intercept overloaded trucks. Optimization of dynamic overload control vehicle patrol routes based on demand forecasting can help improve the deterrent effect of highway overload control and reduce overloaded transport vehicles on highways. Summary of the Invention

[0006] The purpose of this invention is to provide a method, system, device, and medium for dynamic mobile overload control vehicle route planning based on demand forecasting. This method first predicts truck traffic flow to determine the selection and replacement of demand points, and then dynamically plans the routes for the overload control vehicles based on the changing demand. This method is beneficial to the operation of existing overload control models and can reduce costs, thereby strengthening the supervision and control of overloaded trucks.

[0007] A method for planning the patrol route of a dynamic mobile vehicle for overloading control based on demand forecasting includes:

[0008] (1) Establish a demand prediction model for overload control to predict the demand for overload control;

[0009] (2) Determine the demand points for overloading control based on the forecast results;

[0010] (3) Plan the routes of mobile overload control vehicles according to the needs of overload control.

[0011] Furthermore, the establishment of the overload control demand prediction model for predicting overload control demand is detailed below:

[0012] (1) Processing of historical truck inspection data

[0013] Before making predictions, historical data on overloading control is processed, including:

[0014] Remove erroneous data, specifically data with a high missing rate (greater than 80%) and low importance (excluding data from oversized trucks);

[0015] Data integration combines and stores oversized truck data from multiple data sources into a consistent data warehouse;

[0016] Data transformation includes normalizing, discretizing, and sparsifying data to make the dimensions of different features in the data consistent.

[0017] (2) Establish a demand prediction model for overloading control

[0018] The aforementioned demand prediction model for overloading control is a time network model, which is designed with three gating units: input gate, output gate and forget gate, including equations (1)-(6).

[0019] Represents the Gate of Oblivion; Represents the input gate; Representing the output gates, the calculation formulas for each gate are as follows:

[0020] (1)

[0021] (2)

[0022] (3)

[0023] (4)

[0024] (5)

[0025] (6)

[0026] In the formula:

[0027] This represents the element-wise multiplication of two vectors at corresponding positions.

[0028] This represents the sigmoid function, used as an activation function.

[0029] Indicates the current state of the input unit. Represents the cell state vector. Output vector for the current layer;

[0030] Represents the weight matrix. Represents the bias vector;

[0031] Represents the input vector. This represents the output vector of the previous layer.

[0032] A truck spatiotemporal feature matrix containing time and spatial information is constructed as the input to the overload control demand prediction model. After processing, the truck spatiotemporal feature matrix is ​​input into the LSTM (Long Short-Term Memory) layer to output the predicted truck traffic at the next time step.

[0033] The construction of the spatiotemporal feature matrix of trucks is to divide a workday Divided into The time granularity is as follows: the horizontal axis represents the traffic data sorted chronologically, and the vertical axis represents the road segment number, as shown in Equation (7):

[0034] (7)

[0035] in:

[0036] It is a section of road exist Real-time truck traffic data;

[0037] It indicates a road segment The former Historical truck traffic data at any given moment.

[0038] Furthermore, the specific details of determining the overload control demand points based on the forecast results are as follows:

[0039] (1) Determine the static overload control demand points based on the predicted truck traffic volume.

[0040] No. The number of trucks inspected on this road section is determined by the matrix. express. For road section The number of overloaded trucks inspected. For road section Detect the number of trucks. To determine the over-limit rate of road sections, a threshold is set for the over-limit rate of road sections with high over-limit rates. Road sections with an overload rate exceeding the threshold are designated as high-incidence overload sections and are set as static overload control demand points. The judgment formula is as shown in equation (8):

[0041] (8)

[0042] In the formula: Indicate whether additional static overload control points are needed in the overload control area. This indicates that additional areas need to be established for controlling overloading. This indicates that no additional setup is needed.

[0043] (2) Determine the dynamic overload control demand points based on the detected and predicted truck traffic flow.

[0044] Calculate the probability of adding dynamic overload control demand points in each area. As in equation (9):

[0045] (9)

[0046] In the formula:

[0047] This represents the variance of the total number of inspections in the area designated for controlling overloading.

[0048] This represents the minimum variance of the total historical detection volume;

[0049] This represents the maximum variance of the total historical detection volume;

[0050] Set probability threshold Assess and predict truck traffic flow on high-incidence road sections with overloaded and oversized vehicles, in addition to static overload control demand points. If dynamic overload control demand points are added to the overload control area, the probability of overload control will be considered. Less than the probability threshold Therefore, this section of road with a high incidence of overloading and exceeding limits is a new dynamic demand point for overloading control, and the judgment formula is as shown in equation (10):

[0051] (10)

[0052] In the formula: This indicates whether dynamic overload control demand points need to be added in the overload control area. This indicates that additional areas need to be established for controlling overloading. This indicates that no additional setup is needed.

[0053] Furthermore, the step of planning the route of the mobile overload control vehicle based on the demand points for overload control includes establishing a patrol route planning model for the overload control vehicle, the specific content of which is as follows:

[0054] Given the sets of vehicle inspection stations and inspection points, the goal is to enable vehicle inspection vehicles to patrol and inspect all inspection points with minimal cost. The input to the vehicle inspection vehicle patrol route planning model is the vehicle inspection station and the inspection points; the output is the route and order in which the vehicle patrols each inspection point from the inspection station. The dynamic programming process is as follows: A five-working-day period is used as a cycle; within each cycle, the time slice of the previous working day is used... Under the premise that the path is effective, the path of the overloaded vehicle in the next time slice is dynamically planned. This dynamic planning is repeated until the end of the cycle. The goal is to minimize the patrol distance cost and time window penalty cost of the overloaded vehicle. The objective function is Equation (11) and (12), and the constraints are Equation (13)-(21). The overloaded vehicle patrol path planning model is as shown in Equation (11)-(21).

[0055] (11)

[0056] (12)

[0057] ST

[0058] (13)

[0059] (14)

[0060] (15)

[0061] (16)

[0062] (17)

[0063] (18)

[0064] (19)

[0065] (20)

[0066] (twenty one)

[0067] In the formula:

[0068] Indicates the first Under the condition of time slices The cost of a time slice;

[0069] This represents the time window penalty function;

[0070] This indicates a collection of overloaded vehicles, indicated by the subscript. express;

[0071] Represents the set of demand points, with subscripts using express;

[0072] This indicates that overtaking is controlled within the nth time slice. From the demand point arrive When it is 1, it is 0; otherwise, it is 0.

[0073] This indicates that overtaking is controlled within the nth time slice. Do patrols pass through the demand points from the overload control station? The value is 1 when the time has elapsed, and 0 otherwise.

[0074] This represents an intermediate node within the set of demand points.

[0075] This indicates that overtaking is controlled within the nth time slice. From the perspective of demand Patrol through intermediate nodes The value is 1 when the time has elapsed, and 0 otherwise.

[0076] This indicates that overtaking is controlled within the nth time slice. From intermediate node Patrols passing through demand points The value is 1 when the time has elapsed, and 0 otherwise.

[0077] The cost of patrolling by vehicles used for weight control is expressed in yuan per hour per kilometer.

[0078] Indicates the first Cost of overloading control per time slice, unit: yuan / hour·km;

[0079] express Distance between two points, unit: kilometers;

[0080] This indicates the maximum patrol time that the overloaded vehicle can take, in hours;

[0081] Indicates overloaded vehicles The number of points required along the parade route, in units of points;

[0082] Indicates the time when the overloaded vehicle departs from the overload control station;

[0083] Indicates that the overloaded vehicle is from Drive to Time, in hours;

[0084] The speed of the patrol vehicle used for weight control is expressed in kilometers per hour.

[0085] Indicates the first Each time slice represents the detection time of overloaded vehicles at the required points, in hours;

[0086] Indicates the first A time slice, controlling overtaking vehicles From the overload control station to the demand point The required time, of which Unit: hour;

[0087] Indicates the first A time slice, controlling overtaking vehicles From the vehicle control station to the point of demand Time required to begin testing, in hours;

[0088] Represents a maximum value;

[0089] Indicate demand points Maximum cruise time window;

[0090] This represents the time window penalty function;

[0091] Indicate demand points Lower limit of the parade time window.

[0092] Under constraints:

[0093] Formula (13) indicates that the overload control demand point within each workday time slice is only allowed to be patrolled by a single overload control vehicle.

[0094] Formula (14) represents the constraint that the patrol route forms a closed loop after the patrol vehicle returns to the fixed overload control station after each working day time slice.

[0095] Formula (15) ensures that the patrol time of each overloaded vehicle during the workday does not exceed its maximum patrol time;

[0096] Formula (16) is the formula for calculating the number of demand points;

[0097] Formula (17) constrains the initial value of the departure time of the overloaded vehicle, that is, the time when the overloaded vehicle departs from the overloaded station is 0;

[0098] Formula (18) lists the formula for calculating the travel time of the overloaded vehicle, that is, the time is equal to the distance between the two points divided by the speed of the overloaded vehicle;

[0099] Formula (19) constrains the vehicle inspection time to be unaffected by other uncertainties, i.e., the time from the overload control station to the demand point. Required time, the point of demand Detection time and from Arrive at the next demand point The total time required, including all three factors, is less than or equal to the time required for overloading. From the vehicle control station to the point of demand Time required to begin testing;

[0100] Formula (20) constrains the decision variables to be 0-1 variables;

[0101] Formula (21) represents the time window penalty cost constraint for overloaded vehicles.

[0102] A dynamic mobile overload control vehicle patrol route planning system based on demand forecasting, the route planning system is used to implement the above-mentioned dynamic mobile overload control vehicle patrol route planning method based on demand forecasting, including an overload control demand forecasting module, an overload control demand point determination module, and a mobile overload control vehicle route planning module.

[0103] The overload control demand prediction module is used to establish an overload control demand prediction model to predict the overload control demand.

[0104] The module for determining the overload control demand point is used to determine the overload control demand point based on the overload control demand prediction results.

[0105] The module for planning the route of the mobile overload control vehicle is used to plan the route of the mobile overload control vehicle according to the overload control demand points.

[0106] Furthermore, the overload control demand prediction module includes a historical truck data processing unit and an overload control demand prediction model unit;

[0107] The historical truck data processing unit is used to process historical overload control data before prediction.

[0108] The unit for establishing a demand prediction model for overload control is used to establish a demand prediction model for overload control.

[0109] Furthermore, the module for determining the overload control demand points includes a unit for determining static overload control demand points and a unit for determining dynamic overload control demand points;

[0110] The unit for determining static overload control demand points is used to determine static overload control demand points based on the predicted truck traffic flow.

[0111] The unit for determining dynamic overload control demand points is used to determine dynamic overload control demand points based on detected and predicted truck traffic flow.

[0112] Furthermore, the module for planning the route of the mobile overload control vehicle includes a unit for establishing a planning model for the patrol route of the overload control vehicle;

[0113] The unit for establishing the patrol route planning model for overloaded vehicles is used to establish the patrol route planning model for overloaded vehicles.

[0114] An apparatus characterized in that it comprises one or more processors;

[0115] Memory, used to store one or more programs;

[0116] When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described above.

[0117] A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method described above.

[0118] The advantages of this invention compared to existing technologies are as follows:

[0119] The method of this invention improves upon the current approach of controlling overloaded and oversized vehicles, which mainly relies on fixed overload control stations. It effectively prevents overloaded trucks from detouring to avoid overload control stations, improves the efficiency and effectiveness of overload control, reduces costs, and minimizes the harm caused by overloaded and oversized vehicles. Attached Figure Description

[0120] Figure 1 Schematic diagram of the highway network in the overload control area of ​​this invention;

[0121] Figure 2 The truck traffic flow map of the overload control area in this invention;

[0122] Figure 3 The truck traffic flow map predicted for road sections in this invention;

[0123] Figure 4Spatial distribution and distance matrix of static demand points in this invention;

[0124] Figure 5 In this invention, [0, [During the period, a map showing the patrol route for vehicles subject to traffic enforcement was provided.]

[0125] Figure 6 In this invention [ , [Overtaking and parade route map during the period;]

[0126] Figure 7 This is a flowchart of a dynamic mobile overload control vehicle patrol route planning method based on demand forecasting, as described in this invention. Detailed Implementation

[0127] The present invention provides a specific embodiment of a dynamic mobile overload control vehicle patrol route planning method based on demand prediction, which gives the implementation process and solution results. However, the scope of protection of the present invention is not limited to the following embodiment.

[0128] See Figure 7 A method for dynamic mobile overload control vehicle patrol route planning based on demand forecasting, comprising:

[0129] (1) Establish a demand prediction model for overload control to predict the demand for overload control;

[0130] (2) Determine the demand points for overloading control based on the forecast results;

[0131] (3) Plan the routes of mobile overload control vehicles according to the needs of overload control.

[0132] The establishment of a demand prediction model for controlling overloading is described in detail below:

[0133] (1) Processing of historical truck inspection data

[0134] Before making predictions, historical data on overloading control is processed, including:

[0135] Remove erroneous data, specifically data with a high missing rate (greater than 80%) and low importance (excluding data from oversized trucks);

[0136] Data integration combines and stores oversized truck data from multiple data sources into a consistent data warehouse;

[0137] Data transformation includes normalizing, discretizing, and sparsifying data to make the dimensions of different features in the data consistent.

[0138] (2) Establish a demand prediction model for overloading control

[0139] The aforementioned demand prediction model for overloading control is a time network model, which is designed with three gating units: input gate, output gate and forget gate, including equations (1)-(6).

[0140] Represents the Gate of Oblivion; Represents the input gate; Representing the output gates, the calculation formulas for each gate are as follows:

[0141] (1)

[0142] (2)

[0143] (3)

[0144] (4)

[0145] (5)

[0146] (6)

[0147] In the formula:

[0148] This represents the element-wise multiplication of two vectors at corresponding positions.

[0149] This represents the sigmoid function, used as an activation function.

[0150] Indicates the current state of the input unit. Represents the cell state vector. Output vector for the current layer;

[0151] Represents the weight matrix. Represents the bias vector;

[0152] Represents the input vector. This represents the output vector of the previous layer.

[0153] A truck spatiotemporal feature matrix containing time and spatial information is constructed as the input to the overload control demand prediction model. After processing, the truck spatiotemporal feature matrix is ​​input into the LSTM (Long Short-Term Memory) layer to output the predicted truck traffic at the next time step.

[0154] The construction of the spatiotemporal feature matrix of trucks is to divide a workday Divided into The time granularity is as follows: the horizontal axis represents the traffic data sorted chronologically, and the vertical axis represents the road segment number, as shown in Equation (7):

[0155] (7)

[0156] in:

[0157] It is a section of road exist Real-time truck traffic data;

[0158] It indicates a road segment The former Historical truck traffic data at any given moment.

[0159] The method for planning the patrol route of a dynamic mobile vehicle for overload control based on demand forecasting is characterized by the following: the determination of overload control demand points based on the overload control demand forecasting results is detailed below:

[0160] (2) Determine the static overload control demand points based on the predicted truck traffic volume.

[0161] No. The number of trucks inspected on this road section is determined by the matrix. express. For road section The number of overloaded trucks inspected. For road section Detect the number of trucks. To determine the over-limit rate of road sections, a threshold is set for the over-limit rate of road sections with high over-limit rates. Road sections with an overload rate exceeding the threshold are designated as high-incidence overload sections and are set as static overload control demand points. The judgment formula is as shown in equation (8):

[0162] (8)

[0163] In the formula: Indicate whether additional static overload control points are needed in the overload control area. This indicates that additional areas need to be established for controlling overloading. This indicates that no additional setup is needed.

[0164] (2) Determine the dynamic overload control demand points based on the detected and predicted truck traffic flow.

[0165] Calculate the probability of adding dynamic overload control demand points in each area. As in equation (9):

[0166] (9)

[0167] In the formula:

[0168] This represents the variance of the total number of inspections in the area designated for controlling overloading.

[0169] This represents the minimum variance of the total historical detection volume;

[0170] This represents the maximum variance of the total historical detection volume;

[0171] Set probability threshold Assess and predict truck traffic flow on high-incidence road sections with overloaded and oversized vehicles, in addition to static overload control demand points. If dynamic overload control demand points are added to the overload control area, the probability of overload control will be considered. Less than the probability threshold Therefore, this section of road with a high incidence of overloading and exceeding limits is a new dynamic demand point for overloading control, and the judgment formula is as shown in equation (10):

[0172] (10)

[0173] In the formula: This indicates whether dynamic overload control demand points need to be added in the overload control area. This indicates that additional areas need to be established for controlling overloading. This indicates that no additional setup is needed.

[0174] The planning of mobile overload control vehicle routes based on the demand points for overload control includes establishing a patrol route planning model for overload control vehicles, the specific content of which is as follows:

[0175] Given the sets of vehicle inspection stations and inspection points, the goal is to enable vehicle inspection vehicles to patrol and inspect all inspection points with minimal cost. The input to the vehicle inspection vehicle patrol route planning model is the vehicle inspection station and the inspection points; the output is the route and order in which the vehicle patrols each inspection point from the inspection station. The dynamic programming process is as follows: A five-working-day period is used as a cycle; within each cycle, the time slice of the previous working day is used... Under the premise that the path is effective, the path of the overloaded vehicle in the next time slice is dynamically planned. This dynamic planning is repeated until the end of the cycle. The goal is to minimize the patrol distance cost and time window penalty cost of the overloaded vehicle. The objective function is Equation (11) and (12), and the constraints are Equation (13)-(21). The overloaded vehicle patrol path planning model is as shown in Equation (11)-(21).

[0176] (11)

[0177] (12)

[0178] TT

[0179] (13)

[0180] (14)

[0181] (15)

[0182] (16)

[0183] (17)

[0184] (18)

[0185] (19)

[0186] (20)

[0187] (twenty one)

[0188] In the formula:

[0189] Indicates the first Under the condition of time slices The cost of a time slice;

[0190] This represents the time window penalty function;

[0191] This indicates a collection of overloaded vehicles, indicated by the subscript. express;

[0192] Represents the set of demand points, with subscripts using express;

[0193] This indicates that overtaking is controlled within the nth time slice. From the demand point arrive When it is 1, it is 0; otherwise, it is 0.

[0194] This indicates that overtaking is controlled within the nth time slice. Do patrols pass through the demand points from the overload control station? The value is 1 when the time has elapsed, and 0 otherwise.

[0195] This represents an intermediate node within the set of demand points.

[0196] This indicates that overtaking is controlled within the nth time slice. From the perspective of demand Patrol through intermediate nodes The value is 1 when the time has elapsed, and 0 otherwise.

[0197] This indicates that overtaking is controlled within the nth time slice. From intermediate node Patrols passing through demand points The value is 1 when the time has elapsed, and 0 otherwise.

[0198] The cost of patrolling by vehicles used for weight control is expressed in yuan per hour per kilometer.

[0199] Indicates the first Cost of overloading control per time slice, unit: yuan / hour·km;

[0200] express Distance between two points, unit: kilometers;

[0201] This indicates the maximum patrol time that the overloaded vehicle can take, in hours;

[0202] Indicates overloaded vehicles The number of points required along the parade route, in units of points;

[0203] Indicates the time when the overloaded vehicle departs from the overload control station;

[0204] Indicates that the overloaded vehicle is from Drive to Time, in hours;

[0205] The speed of the patrol vehicle used for weight control is expressed in kilometers per hour.

[0206] Indicates the first Each time slice represents the detection time of overloaded vehicles at the required points, in hours;

[0207] Indicates the first A time slice, controlling overtaking vehicles From the overload control station to the demand point The required time, of which Unit: hour;

[0208] Indicates the first A time slice, controlling overtaking vehicles From the vehicle control station to the point of demand Time required to begin testing, in hours;

[0209] Represents a maximum value;

[0210] Indicate demand points Maximum cruise time window;

[0211] This represents the time window penalty function;

[0212] Indicate demand points Lower limit of the parade time window.

[0213] Under constraints:

[0214] Formula (13) indicates that the overload control demand point within each workday time slice is only allowed to be patrolled by a single overload control vehicle.

[0215] Formula (14) represents the constraint that the patrol route forms a closed loop after the patrol vehicle returns to the fixed overload control station after each working day time slice.

[0216] Formula (15) ensures that the patrol time of each overloaded vehicle during the workday does not exceed its maximum patrol time;

[0217] Formula (16) is the formula for calculating the number of demand points;

[0218] Formula (17) constrains the initial value of the departure time of the overloaded vehicle, that is, the time when the overloaded vehicle departs from the overloaded station is 0;

[0219] Formula (18) lists the formula for calculating the travel time of the overloaded vehicle, that is, the time is equal to the distance between the two points divided by the speed of the overloaded vehicle;

[0220] Formula (19) constrains the vehicle inspection time to be unaffected by other uncertainties, i.e., the time from the overload control station to the demand point. Required time, the point of demand Detection time and from Arrive at the next demand point The total time required, including all three factors, is less than or equal to the time required for overloading. From the vehicle control station to the point of demand Time required to begin testing;

[0221] Formula (20) constrains the decision variables to be 0-1 variables;

[0222] Formula (21) represents the time window penalty cost constraint for overloaded vehicles.

[0223] A dynamic mobile overload control vehicle patrol route planning system based on demand forecasting, the route planning system is used to implement the above-mentioned dynamic mobile overload control vehicle patrol route planning method based on demand forecasting, including an overload control demand forecasting module, an overload control demand point determination module, and a mobile overload control vehicle route planning module.

[0224] The overload control demand prediction module is used to establish an overload control demand prediction model to predict the overload control demand.

[0225] The module for determining the overload control demand point is used to determine the overload control demand point based on the overload control demand prediction results.

[0226] The module for planning the route of the mobile overload control vehicle is used to plan the route of the mobile overload control vehicle according to the overload control demand points.

[0227] Furthermore, the overload control demand prediction module includes a historical truck data processing unit and an overload control demand prediction model unit;

[0228] The historical truck data processing unit is used to process historical overload control data before prediction.

[0229] The unit for establishing a demand prediction model for overload control is used to establish a demand prediction model for overload control.

[0230] Furthermore, the module for determining the overload control demand points includes a unit for determining static overload control demand points and a unit for determining dynamic overload control demand points;

[0231] The unit for determining static overload control demand points is used to determine static overload control demand points based on the predicted truck traffic flow.

[0232] The unit for determining dynamic overload control demand points is used to determine dynamic overload control demand points based on detected and predicted truck traffic flow.

[0233] Furthermore, the module for planning the route of the mobile overload control vehicle includes a unit for establishing a planning model for the patrol route of the overload control vehicle;

[0234] The unit for establishing the patrol route planning model for overloaded vehicles is used to establish the patrol route planning model for overloaded vehicles.

[0235] An apparatus characterized in that it comprises one or more processors;

[0236] Memory, used to store one or more programs;

[0237] When the one or more programs are executed by the one or more processors, the one or more processors perform the method as described above.

[0238] A computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method described above.

[0239] Based on the aforementioned method for planning the patrol route of a dynamic mobile overload control vehicle based on demand forecasting, this invention provides another device. The device includes, but is not limited to, one or more processors and a memory.

[0240] Memory, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and modules, such as the program instructions corresponding to the dynamic mobile overload control vehicle patrol route planning method based on demand prediction in this embodiment of the invention. The processor executes the software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the vehicle, thereby realizing the aforementioned dynamic mobile overload control vehicle patrol route planning method based on demand prediction.

[0241] The memory may primarily include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the terminal. Furthermore, the memory may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0242] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for planning the patrol route of a dynamic mobile overload control vehicle based on demand prediction. This method includes the following steps:

[0243] First, establish a demand prediction model for controlling overloading to predict the demand for overloading control;

[0244] Secondly, the demand points for controlling overloading are determined based on the forecast results of the demand for controlling overloading.

[0245] Secondly, the routes for mobile overload control vehicles should be planned based on the needs of overload control.

[0246] The computer-readable storage medium provided by the present invention has computer-executable instructions that are not limited to the method operations described above, but can also execute related operations in the dynamic mobile overload control vehicle patrol path planning method based on demand prediction provided in any embodiment of the present invention.

[0247] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this application can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0248] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line DSL) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk, SSD), etc.

[0249] Based on the predicted dynamic demand for overload control, the route planning of mobile overload control vehicles is an NP-hard problem that requires a heuristic algorithm to solve. This invention uses the ant colony algorithm to solve the overload control vehicle patrol route planning model, as detailed below:

[0250] (1) Path construction

[0251] A virtual starting point is set, from which all ants begin. The transition probability of a first-level ant choosing each feasible node is determined by the selection expectation between the current node and feasible nodes, and the pheromone level. Based on this principle, the ant transition process and pheromone update method in the ant colony algorithm are designed.

[0252] The method is as shown in equations (22)-(24):

[0253] (twenty two)

[0254] (twenty three)

[0255] (24)

[0256] In the formula: 、 are the pheromone concentration and the expected heuristic factor respectively; is the pheromone concentration of route ab at time T when optimizing the location points of H overloading control stations;

[0257] (2) Solution steps

[0258] 1) Initialize parameters At the beginning of the calculation, relevant parameters need to be initialized, input the total number N of overloading control demand points, the maximum cruising time of each overloading control vehicle

[0259] 2) Construct the solution space Randomly place each ant at different starting points, and calculate the next demand point to be visited for each ant until all ants have visited all demand points.

[0260] 3) Update pheromone Calculate the path length passed by each ant, and record the optimal solution (the shortest path) in the current iteration. At the same time, update the pheromone concentration on the connection paths of each demand point.

[0261] 4) Judge whether to terminate If iter < iter_max, then set iter = iter + 1, clear the record table of the paths passed by the ants, and return to step 2; otherwise, terminate the calculation and output the optimal solution.

[0262] 5) Judge whether to terminate If iter < iter_max, then set iter = iter + 1, clear the record table of the paths passed by the ants, and return to step 2; otherwise, terminate the calculation and output the optimal solution. 3. Judge whether to terminate If iter < iter_max, then set iter = iter + 1, clear the record table of the paths passed by the ants, and return to step 2; otherwise, terminate the calculation and output the optimal solution.

[0263] Embodiment

[0264] 1. The method described in the present invention is applied to the overloading control area of Songyuan area, Jilin Province. As Figure 1 shown.

[0265] 2. Processing of historical detected truck data

[0266] Process the data of overloading control points and toll stations near Xiaojia Overloading Control Station in Songyuan from July 27 to October 27, 2022 for a total of 3 months, delete 641 pieces of data, and the remaining 8779 pieces of data. Divide the remaining data, and divide the data set at a granularity of 30 minutes. The granularity of 30 minutes can leave enough buffer time for real-time adjustment decisions in actual work management. After division, the data set of one day is divided into 48 data sets.

[0267] 3. Forecast the demand for overloading control.

[0268] A spatial distribution map of the planned patrol road network within a 40km radius of the overload control station was constructed, and 32 road segments within the area were numbered. The spatial distribution of the road network in the study area is as follows: Figure 1 As shown in the image, yellow represents highways. A section within the designated area for controlling overloaded vehicles was selected to predict truck traffic flow. After data filtering and segmentation, the truck traffic flow for this section was calculated every 30 minutes. These 30-minute traffic flows for all data types were then summed to obtain the total truck traffic flow for the section. The daily distribution of truck traffic flow on July 30th is shown below. Figure 2 .

[0269] The processed data was input into the prediction model in the form of a traffic sequence as shown in Equation (7) to predict truck traffic flow. The prediction model is shown in Equations (1)-(6). The model was trained using data from the first two months (July 27, 2023 - September 27, 2023) and validated using data from the third month (September 28, 2023 - October 27, 2023). After parameter adjustment, the maximum number of iterations was determined to be 500, and the traffic flow of the previous four days was used to predict the traffic flow of the next day, with a learning rate of 0.01. The RMSE of the prediction result on the test set was 0.27813. The truck traffic flow predicted for segment 19 is as follows: Figure 3 As shown in the figure (truck traffic unit: vehicles, 1 time point represents 30 minutes),

[0270] 4. Determine the demand points for overloading control based on the forecast results.

[0271] The study assesses the overload performance of road sections based on predicted truck traffic flow, identifies high-incidence road sections and times for overload, and sets the overload rate of more than 3% as a seriously overloaded road section, with the high-incidence time of overload obtained from predicted truck time series data.

[0272] [0, Dynamic overload control route optimization within the interval sets the midpoint of the high-incidence overload section as the static demand point, and then calculates the distance matrix between the static demand points. This transforms the problem into a typical path problem, within the range [0, ... Nine static demand points are generated within the interval. Dijkstra's algorithm is used to calculate the distance matrix between the demand points. The spatial distribution of the static demand points and the distance matrix within the interval are shown below. Figure 4 As shown (the data in the table represents the distance between two demand points).

[0273] 5. Plan the routes of mobile overload control vehicles based on the needs of overload control.

[0274] Assuming each vehicle travels at 50 km / h, requires 40 minutes to stop and inspect at a single inspection point, has a maximum possible patrol time of 5 hours, and a fuel cost of 0.6 yuan per kilometer per vehicle, then... Nine static demand points are generated within the interval. These are substituted into the overload control vehicle patrol route planning model, as shown in equations (11)-(21). The ant colony algorithm is used to solve the problem, as shown in equations (22)-(24). The obtained overload control vehicle route scheme at t=0 is shown in the figure. The results are as follows: Figure 5 As shown (points of the same shape represent demand points on the same route), at time t=0, the overload control station dispatched a total of 3 overload control vehicles to [0, During the period, patrols were conducted at 9 static overload control points, with routes 1: 1->2->7->8->1, 2: 1->3->5->4->1, and 3: 1->9->6->10->1, covering a total distance of 203.41km and costing 136.452 yuan.

[0275] The patrol detection data, in the form of a traffic sequence as shown in equation (7), is then fed into the prediction model for truck traffic prediction. The prediction model is shown in equations (1)-(6). Then, dynamic demand points are determined using equations (8), (9), and (10). New dynamic demand points are determined to dynamically update the patrol plan for overloaded vehicles. If no new dynamic demand points are generated, the original patrol route continues. , During this period, three new dynamic demand points were generated. Furthermore, due to the detours of trucks that bypassed the sections frequently patrolled by the vehicle enforcement patrol, the number of trucks detected at the original two static demand points (3 and 8) decreased. The updated demand points were then incorporated into the vehicle enforcement patrol route planning model, as shown in equations (11)-(21), and solved using the ant colony algorithm, as shown in equations (22)-(24). Therefore, the vehicle enforcement patrol plan needs to be updated as follows: Figure 6 As shown (points of the same shape represent demand points on the same route). The results indicate that in [ , At that moment, the overload control station dispatched three overload control vehicles to [0, During the period, patrols were conducted at 9 static overload control points, with routes 1: 1->9->6->13->1, 2: 1->12->5->4->1, and 3: 1->10->2->7->11->1, for a total distance of 186.09km and a total cost of 124.846 yuan.

[0276] The working efficiency W and utilization rate UR of the overload control vehicle can be expressed by the following formulas:

[0277] (25)

[0278] (26)

[0279] In the formula:

[0280] For the required points;

[0281] The actual patrol time for vehicle K, used for controlling overtaking, is expressed in hours.

[0282] A collection of overload control vehicles assigned to all overload control stations;

[0283] The maximum patrol time for each vehicle used for weight enforcement is given in hours.

[0284] From equations (25) and (26), we can obtain [0, The overall work efficiency of overtaking within the time slice was 84.91%, and the utilization rate was 70.66%. , At that time, the overall work efficiency of the overload control vehicles was 92.76%, and the utilization rate was 71.87%.

[0285] In summary, overload control vehicles can effectively intercept trucks, with a utilization rate exceeding 80% and good operational efficiency. This verifies the feasibility of the model of this invention in optimizing overload control vehicle routes, which is beneficial to the operation of existing overload control models, can reduce costs, strengthen the supervision and control of overloaded trucks, and can help managers implement better route decisions for overload control vehicles.

[0286] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be included within the scope of protection of the present invention. Furthermore, all content not described in detail in this specification is prior art known to those skilled in the art.

Claims

1. A method for planning the patrol route of a dynamic mobile overload control vehicle based on demand forecasting, comprising: (1) Establish a demand prediction model for overload control to predict the demand for overload control; (2) Determine the demand points for overloading control based on the forecast results; (3) Plan the routes of mobile overload control vehicles according to the needs of overload control; The specific details of determining the demand points for overloading control based on the forecast results are as follows: Determine the static overload control demand points based on the predicted truck traffic flow; No. The number of trucks inspected on this road section is determined by the matrix. express; For road section The number of overloaded trucks inspected. For road section Detect the number of trucks. To determine the over-limit rate of road sections, a threshold is set for the over-limit rate of road sections with high over-limit rates. Road sections with an over-limit rate greater than the threshold are designated as high-incidence over-limit road sections and are set as static over-limit control demand points; the judgment formula is as shown in equation (8): (8) In the formula: Indicate whether additional static overload control points are needed in the overload control area. This indicates that additional areas need to be established for controlling overloading. This indicates that no additional setup is needed; Determine the dynamic demand points for overloading control based on the detected and predicted truck traffic flow; Calculate the probability of adding dynamic overload control demand points in each area. As in equation (9): (9) In the formula: This represents the variance of the total number of inspections in the area designated for controlling overloading. This represents the minimum variance of the total historical detection volume; This represents the maximum variance of the total historical detection volume; Set probability threshold Assess and predict truck traffic flow on high-incidence road sections with overloaded and oversized vehicles, in addition to static overload control demand points. If dynamic overload control demand points are added to the overload control area, the probability of overload control will be considered. Less than the probability threshold Therefore, this section of road with a high incidence of overloading and exceeding limits is a new dynamic demand point for overloading control, and the judgment formula is as shown in equation (10): (10) In the formula: This indicates whether dynamic overload control demand points need to be added in the overload control area. This indicates that additional areas need to be established for controlling overloading. This indicates that no additional setup is needed; The planning of mobile overload control vehicle routes based on the demand points for overload control includes establishing a patrol route planning model for overload control vehicles, the specific content of which is as follows: Given the known set of overload control stations and overload control demand points, the goal is to enable overload control vehicles to patrol and inspect all overload control demand points at the lowest cost. The input to the overload control vehicle patrol route planning model is the overload control station and the overload control demand point, and the output is the route and order in which the overload control vehicle patrols each demand point from the overload control station. The dynamic programming process is as follows: A five-working-day period is defined as the time slice of the previous working day. Under the premise that the path is effective, the path of the overloaded vehicle in the next time slice is dynamically planned, and this dynamic planning is repeated until the end of the cycle, with the goal of minimizing the patrol distance cost and time window penalty cost of the overloaded vehicle.

2. The method for dynamic mobile overload control vehicle patrol route planning based on demand forecasting according to claim 1, characterized in that: The establishment of a demand prediction model for controlling overloading is described in detail below: (1) Processing of historical truck inspection data Before making predictions, historical data on overloading control is processed, including: Remove erroneous data, that is, remove data with a high missing rate and low importance; Data integration combines and stores oversized truck data from multiple data sources into a consistent data warehouse; Data transformation includes normalizing, discretizing, and sparsifying data to make the dimensions of different features in the data consistent. (2) Establish a demand prediction model for overloading control The aforementioned demand prediction model for overloading control is a time network model, which is designed with three gate control units: input gate, output gate and forget gate, including equations (1)-(6). Represents the Gate of Oblivion; Represents the input gate; Representing the output gates, the calculation formulas for each gate are as follows: (1) (2) (3) (4) (5) (6) In the formula: This represents the element-wise multiplication of two vectors at corresponding positions. This represents the sigmoid function, used as an activation function. Indicates the current state of the input unit. Represents the cell state vector. Output vector for the current layer; Represents the weight matrix. Represents the bias vector; Represents the input vector. This represents the output vector of the previous layer; A truck spatiotemporal feature matrix containing time and space information is constructed as the input to the overload control demand prediction model. After processing, the truck spatiotemporal feature matrix is ​​input into the LSTM layer to output the predicted truck traffic at the next time step. The construction of the spatiotemporal feature matrix of trucks is to divide a workday Divided into The time granularity is as follows: the horizontal axis represents the traffic data sorted chronologically, and the vertical axis represents the road segment number, as shown in Equation (7): (7) in: It is a section of road exist Real-time truck traffic data; It indicates a road segment The former Historical truck traffic data at any given moment.

3. The method for dynamic mobile overload control vehicle patrol route planning based on demand forecasting according to claim 1, characterized in that: The objective functions are equations (11) and (12), and the constraints are equations (13)-(21). The overloaded vehicle patrol route planning model is shown in equations (11)-(21). (11) (12) S.T. (13) (14) (15) (16) (17) (18) (19) (20) (21) In the formula: Indicates the first Under the condition of time slices The cost of a time slice; This represents the time window penalty function; This indicates a collection of overloaded vehicles, indicated by the subscript. express; Represents the set of demand points, with subscripts using express; This indicates that overtaking is controlled within the nth time slice. From the demand point arrive When it is 1, it is 0; otherwise, it is 0. This indicates that overtaking is controlled within the nth time slice. Do patrols pass through the demand points from the overload control station? The value is 1 when the time has elapsed, and 0 otherwise. This represents an intermediate node within the set of demand points. This indicates that overtaking is controlled within the nth time slice. From the perspective of demand Patrol through intermediate nodes The value is 1 when the time has elapsed, and 0 otherwise. This indicates that overtaking is controlled within the nth time slice. From intermediate node Patrols passing through demand points The value is 1 when the time has elapsed, and 0 otherwise. The cost of patrolling by vehicles used for weight control is expressed in yuan per hour per kilometer. Indicates the first Cost of overloading control per time slice, unit: yuan / hour·km; express Distance between two points, unit: kilometers; This indicates the maximum patrol time that the overloaded vehicle can take, in hours; Indicates overloaded vehicles The number of points required along the parade route, in units of points; Indicates the time when the overloaded vehicle departs from the overload control station; Indicates that the overloaded vehicle is from Drive to Time, in hours; The speed of the patrol vehicle used for weight control is expressed in kilometers per hour. Indicates the first Each time slice represents the detection time of overloaded vehicles at the required points, in hours; Indicates the first A time slice, controlling overtaking vehicles From the overload control station to the demand point The required time, of which Unit: hour; Indicates the first A time slice, controlling overtaking vehicles From the vehicle control station to the point of demand Time required to begin testing, in hours; Represents a maximum value; Indicate demand points Maximum cruise time window; This represents the time window penalty function; Indicate demand points Lower limit of the parade time window; Under constraints: Formula (13) indicates that the overload control demand point within each workday time slice is only allowed to be patrolled by a single overload control vehicle. Formula (14) represents the constraint that the patrol route forms a closed loop after the patrol vehicle returns to the fixed overload control station after each working day time slice. Formula (15) ensures that the patrol time of each overloaded vehicle during the workday does not exceed its maximum patrol time; Formula (16) is the formula for calculating the number of demand points; Formula (17) constrains the initial value of the departure time of the overloaded vehicle, that is, the time when the overloaded vehicle departs from the overloaded station is 0; Formula (18) lists the formula for calculating the travel time of the overloaded vehicle, that is, the time is equal to the distance between the two points divided by the speed of the overloaded vehicle; Formula (19) constrains the vehicle inspection time to be unaffected by other uncertainties, i.e., the time from the overload control station to the demand point. Required time, the point of demand Detection time and from Arrive at the next demand point The total time required, including all three factors, is less than or equal to the time required for overloading. From the vehicle control station to the point of demand Time required to begin testing; Formula (20) constrains the decision variables to be 0-1 variables; Formula (21) represents the time window penalty cost constraint for overloaded vehicles.

4. A dynamic mobile overload control vehicle patrol route planning system based on demand forecasting, characterized in that: The path planning system is used to implement the dynamic mobile overload control vehicle patrol path planning method based on demand prediction as described in any one of claims 1 to 3, including an overload control demand prediction module, an overload control demand point determination module, and a mobile overload control vehicle route planning module. The overload control demand prediction module is used to establish an overload control demand prediction model to predict the overload control demand. The module for determining the overload control demand point is used to determine the overload control demand point based on the overload control demand prediction results. The module for planning the route of the mobile overload control vehicle is used to plan the route of the mobile overload control vehicle according to the overload control demand points.

5. A dynamic mobile overload control vehicle patrol route planning system based on demand forecasting according to claim 4, characterized in that: The overload control demand prediction module includes a historical truck data processing unit and an overload control demand prediction model unit. The historical truck data processing unit is used to process historical overload control data before prediction. The unit for establishing a demand prediction model for overload control is used to establish a demand prediction model for overload control.

6. An apparatus, characterized in that: Includes one or more processors; Memory, used to store one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described in any one of claims 1-3.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by the processor, it implements the method as described in any one of claims 1-3.

Images