Expressway minimum fee path calculation method based on automatic graph partitioning technology

Abstract

The invention belongs to the technical field of intelligent transportation, and provides a method for calculating a minimum fee path of an expressway based on an automatic graph partitioning technology, which comprises the following steps of: constructing an expressway road network model based on charging parameters, determining the number of partitions by combining an actual physical space, and partitioning the road network by adopting the automatic graph partitioning technology to obtain a minimum fee path of the expressway. And fusing the partitions to obtain boundary nodes, calculating block node intermediate parameters, calculating boundary node intermediate parameters, and finally performing combination matching based on mileage amount table parameters in a road section, the block node intermediate parameters and the boundary node intermediate parameters to obtain a minimum fee path between any two toll stations in a road network. Through algorithm optimization, the problem of large memory occupation amount of the intermediate parameters is reduced, the efficiency problem of online charging is considered, and the influence on the calculation efficiency after the intermediate parameters are compressed is reduced.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent transportation, and particularly relates to a calculation method for the minimum toll path of an expressway based on graph automatic partitioning technology. The system provides a method for calculating the expressway passing path of a vehicle using less space and time. Background Art

[0002] At 0:00 on January 1, 2020, the project to cancel the provincial boundary toll stations on expressways was completed and the whole network was switched to grid connection. Before and after the operation of the national "one network", the travel and billing methods of expressways have changed greatly according to the unified technical plan of the Ministry of Transport. Before the cancellation of the provincial boundary toll stations, each province adopted a closed billing and charging mode. Vehicles passing through provinces needed to stop and pay tolls when passing through the provincial boundary toll stations, and the charging systems of each province independently carried out estimated billing; after the cancellation of the provincial boundary toll stations, a segmented billing mode was adopted. Vehicles passing through provinces no longer needed to stop and pay tolls when passing through the provincial boundary, and the exit toll stations could carry out estimated billing through the cross-provincial online billing interface of the central department. In order to implement this billing method, the central department compiled and issued an overall technical plan. First, each province newly built a large number of gantry systems to provide basic equipment support for segmented billing. Secondly, the unified billing module specifications were formulated to restrict the functions, performance, encapsulation, and interfaces of the billing module. Then, each province independently implemented the estimated billing function of the billing module according to the billing module specifications, and then packaged and uploaded it to the central department's rate parameter management system. Finally, the central department uniformly scheduled and managed the billing modules and parameters of each province. On the one hand, the billing modules and parameters were sent to the gantry systems for segmented billing transactions of the gantry systems. On the other hand, the cross-provincial online billing request was initiated to the central department at the toll station lanes in the exit provinces. The central department called the billing modules of each passing province one by one according to the provinces passed by the vehicle to complete the fee calculation and the assembly of the billing results, and finally completed the one-time calculation of the cross-provincial toll.

[0003] In order to calculate the tolls for cross-provincial vehicles, the central department needs to load and call the billing modules across the country. Therefore, in the technical solution of the central department, there are strict requirements for the memory resource occupancy and computing performance of the billing modules in each province. Generally, the path fitting algorithm for vehicles is a greedy algorithm with low computing efficiency. To meet the requirements of computing performance, a common practice is to pre-calculate the intermediate parameters and billing parameters of path fitting during the initialization of the billing module and load them into the memory together, which can greatly improve the computing efficiency of path fitting. However, with the continuous development of expressways, the road network structure of expressways in the province has become increasingly complex, which directly leads to the memory occupancy of the billing module gradually approaching the memory occupancy threshold required by the central department. Among them, the intermediate parameters of path fitting occupy the most memory. Although expanding the memory occupancy threshold can solve the urgent problem, it cannot solve the problem of sustainable development. Moreover, if each province requires an expansion of memory occupancy, the server resources of the central department cannot support unlimited expansion. Therefore, researching how to compress intermediate parameters to achieve path fitting is of great significance to the networked toll collection system.

[0004] However, issues such as how to compress path fitting parameters and how to balance parameter compression and computing performance have become a major research focus in the algorithm for restoring the driving trajectories of highway vehicles. Summary of the Invention

[0005] To solve the technical problem in the prior art that path fitting parameters cannot be effectively compressed, automatic partitioning cannot be achieved, and the minimum toll path of an expressway cannot be quickly obtained through fitting calculation, the present invention proposes a calculation method for the minimum toll path of an expressway based on graph automatic partitioning technology. Automatic partitioning is achieved based on the K-Means++ algorithm. By optimizing the algorithm, the problem of large memory occupancy of intermediate parameters during the calculation of the minimum toll path of an expressway is reduced, the computing efficiency is improved, and the efficiency problem of online billing is also considered.

[0006] The specific technical solution is as follows:

[0007] A calculation method for the minimum toll path of an expressway based on graph automatic partitioning technology,

[0008] S1, constructing an expressway road network model based on billing parameters: Based on the routing node information and the mileage and amount table information within sections in the billing parameters, construct an expressway road network model;

[0009] S2, determining the number of partitions in combination with the actual physical space: Based on the expressway road network model, determine the number of partitions according to the number of traffic hub nodes within the set geographical space;

[0010] S3. Use the graph automatic partitioning technology to partition the road network: adopt the K-Means++ algorithm to automatically divide the hub nodes of the entire road network into K regions, record the region labels for each hub node, and use the silhouette coefficient to evaluate the partitioning quality;

[0011] S4. Integrate the partitions to obtain boundary nodes: generate a set of candidate boundary nodes and a set of boundary lines based on the hub nodes obtained after partitioning in S3, and select boundary nodes and record their affiliated partitions based on the number of times the boundary nodes in the set of candidate boundary nodes appear in the set of boundary lines;

[0012] S5. Calculate the intermediate parameters of block nodes: regard any two hub nodes within the same region as block nodes, construct a hub set of the intermediate parameters of the block nodes, and calculate the intermediate parameters of the minimum-fare path based on the intermediate parameter amounts in the mileage and fare table within the road section as the intermediate parameters of the block nodes;

[0013] S6. Calculate the intermediate parameters of boundary nodes: combine the boundary nodes, the intermediate parameters of the block nodes, and the mileage and fare table within the road section to calculate the intermediate parameters of the minimum-fare path as the intermediate parameters of the boundary nodes;

[0014] S7. Use the intermediate parameters for fast path fitting: combine and match the parameters in the mileage and fare table within the road section, the intermediate parameters of the block nodes, and the intermediate parameters of the boundary nodes to obtain the minimum-fare path between any two toll stations in the road network.

[0015] Preferably, the method for constructing the highway road network model based on the charging parameters in S1 is as follows:

[0016] S11. Aggregate the routing nodes into hub nodes: extract the routing node data from all the charging parameters, aggregate the station codes representing the same routing node together, and take the smallest toll station code as the new hub node code to obtain the set N of all hub nodes in the road network. Store the multiple station codes of the same hub node as the attributes of the hub node as the point information in the road network model;

[0017] S12. Obtain the edge information between hub nodes through the mileage and fare table within the road section: the mileage and fare table within the road section contains the mileage information and fare information between any two stations on the same road section; if two hub nodes are adjacent, directly query the mileage information and fare information from the mileage and fare table within the road section as the edge information; if there are multiple hub nodes on the same road section, there will be records in the mileage and fare table within the road section that cross hub nodes. To obtain the connection relationship between adjacent hub nodes, for the records in the mileage and fare table within the road section with the same starting node, take the record in the mileage and fare table within the road section with the smallest number of toll units passed through to generate the connection relationship of the hub nodes as the edge information;

[0018] S13. Construct a highway road network model based on the point information and edge information.

[0019] Preferably, the method for determining the number of partitions by combining the actual physical space in S2 is as follows: Determine the number of partitions of the entire road network map based on the threshold of the distribution quantity of the hub nodes in the set geographical space, and set the number of partitions as the K value.

[0020] Preferably, the method for partitioning the road network by using the graph automatic partitioning technology in S3 is as follows:

[0021] S31: Use the K-Means++ algorithm to automatically partition the road network: Randomly select a hub node as the first regional center; for the remaining hub nodes, the algorithm calculates the distance between each point and the selected regional center point, and selects the point with the farthest distance as the new regional center point until K regional center points are selected; assign each hub node to the nearest regional center point to form K clusters;

[0022] S32: Use the silhouette coefficient to evaluate the quality of the K clusters: If the silhouette coefficient is less than or equal to 0.5, re-run the K-Means++ algorithm for re-partitioning until the silhouette coefficient is greater than 0.5 and then stop re-partitioning.

[0023] Preferably, the method for calculating the silhouette coefficient is as follows:

[0024] S3A: For each hub node, calculate the average distance between it and all other points in the same cluster and denote it as a;

[0025] S3B: For each hub node, calculate the minimum average distance between it and the points in the nearest cluster other than its own cluster, that is, the average distance from this point to all points in the nearest cluster, and denote it as b;

[0026] S3C: Calculate the silhouette coefficient c of each hub node according to the formula i :

[0027]

[0028] S3D: Calculate the average value of the silhouette coefficients of all hub nodes, which is the silhouette coefficient of the entire cluster.

[0029] Preferably, the method for fusing the partition boundary nodes in step S4 is as follows:

[0030] S41: For each partition, traverse the hub nodes in the partition. If not all the neighbor nodes of a certain hub node are within the partition, record the connection line formed by this node and the neighbor nodes not within the partition as the boundary connection line in the alternative boundary node set;

[0031] S42: Find the edges of all boundary nodes and the boundary nodes of other regions to obtain a set of boundary lines;

[0032] S43: Select the one that appears the most times in the set of boundary lines as the regional boundary point from the set of alternative boundary nodes. If the number of appearances is the same, select the one with a smaller node code as the boundary node, and record the original partition and associated partition information of the boundary node.

[0033] Preferably, the method for finding the minimum fare path in step S5 is as follows:

[0034] S51: Treat any hub node within the same region as a block node. If any two of the block nodes belong to the same road section, obtain the minimum fare path through the mileage and fare table within the road section; if there are multiple road sections, select the one with the smallest fare information as the minimum fare path;

[0035] S52: If a block node is connected to multiple road sections, the block node is regarded as a virtual station, and the minimum fare path of the hub node between two connected road sections is found based on the virtual station code;

[0036] S53: Repeat steps S51 - S52 until the minimum fare path between any two block nodes is found.

[0037] Preferably, the method for calculating the intermediate parameter of the block node in step S5 is as follows:

[0038] S5A: Construct a hub set of intermediate parameters: Treat any hub node within the same region as a block node. For each partition, use the block node and the boundary node associated with the block node together as the hub set S for calculating the intermediate parameter;

[0039] S5B: Initialize the intermediate parameters of any two block nodes: Take the Cartesian product of the nodes in the hub set S, and initialize the intermediate parameters of any two block nodes within the same region and set them as the first queue. If the two combined block nodes are the same, set the fare to 0, otherwise set it to infinity;

[0040] S5C: Put all the intermediate parameters with a fare of 0 into the queue Q1 and regard it as the queue to be allocated;

[0041] S5D: Poll the queue Q1, obtain an intermediate parameter from the head of the queue. Assume the start and end points of the intermediate parameter are a and b, and obtain the data with b as the start point and other adjacent block nodes of b within the region as the end points from the mileage and fare table within the road section to obtain the mileage and fare table set L1 within the road section;

[0042] S5E: Traverse L1. For the data in the mileage amount table within a certain section of the road, assume the starting point and ending point of the data are b and c respectively. If the intermediate parameter amount from a to c is greater than the intermediate parameter amount of other path schemes, that is, the intermediate parameter amount from a to b plus the amount from b to c in the mileage amount table within the section of the road, then update the intermediate parameter amount from a to c to the intermediate parameter amount of other path schemes, and add the intermediate parameter from a to c to the first queue;

[0043] S5F: Loop through the above steps S5A - S5E until all intermediate parameters in the Q1 queue are polled;

[0044] S5G: For the intermediate parameters of the minimum - fee path within the same area in the first queue, filter out the data with an amount of 0 and an infinite - large amount, and obtain the intermediate parameters of the block nodes of the final minimum - fee path.

[0045] Preferably, the method for calculating the intermediate parameters of the boundary nodes in S6 is as follows:

[0046] S61: Combine the boundary nodes by taking the Cartesian product, and initialize the amount of the path to be infinite - large;

[0047] S62: Search for the intermediate parameters of the minimum - fee path within the area, obtain the intermediate parameters of the path between any two boundary nodes within the same area, and update the path information of the boundary nodes;

[0048] S63: For the data with the same boundary node as the starting point and the ending point, if the intermediate parameter amount with the starting point and the ending point of the data as the start and end points is less than the sum of the amounts of the data, then update the path information and amount information of the data;

[0049] S64: Repeat and iterate step S63 until the intermediate parameters of the minimum - fee paths between all boundary nodes are calculated and used as the intermediate parameters of the boundary nodes.

[0050] Preferably, the method for using the intermediate parameters for fast path fitting in S7 is as follows:

[0051] S71: First, through the mileage amount table within the section of the road, find the nearest hub nodes A1, A2 and B1, B2 to the starting toll station and the ending toll station respectively;

[0052] S72: Find the boundary - node sets R1, R2…Rn of the partition to which the starting toll station belongs, and the boundary - node sets R1`, R2`…Rn` of the partition to which the ending toll station belongs;

[0053] S73: Through the starting toll station to the starting hub, from the starting hub to the inner boundary hub of the starting area, from the inner boundary hub of the starting area to the inner boundary hub of the ending area, from the inner boundary hub of the ending area to the ending hub, and from the ending hub to the ending toll station, using a hierarchical connection method. Calculate the Cartesian product of the node sets at each level respectively, search for the minimum toll path information at each level, and finally concatenate them to obtain multiple path information from the starting point to the ending point; Search for the minimum toll path information from the starting toll station to the starting hub and from the ending hub to the ending toll station in the mileage and toll table within the road section; Search for the minimum toll path information from the starting hub to the inner boundary hub of the starting area and from the inner boundary hub of the ending area to the ending hub in the intermediate parameters of the block nodes; Search for the minimum toll path information from the inner boundary hub of the starting area to the inner boundary hub of the ending area in the intermediate parameters of the boundary nodes.

[0054] S74: For multiple toll path information, select the toll path information with the minimum amount as the minimum toll path between any two toll stations in the road network.

[0055] Beneficial effects:

[0056] The present invention proposes a calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology. Build an expressway road network model based on the charging parameters, then determine the number of partitions in combination with the actual physical space, and use the graph automatic partitioning technology to partition the road network. Integrate the partitions to obtain boundary nodes, calculate the intermediate parameters of the block nodes, and calculate the intermediate parameters of the boundary nodes. Finally, perform combined matching based on the mileage and toll table parameters within the road section, the intermediate parameters of the block nodes, and the intermediate parameters of the boundary nodes to obtain the minimum toll path between any two toll stations in the road network. It solves the technical problem that the intermediate parameter algorithm of the original minimum toll path needs to calculate the intermediate parameters of the minimum toll path between the virtual station codes of any two hub nodes in the entire road network, and the intermediate parameters increase exponentially. The present invention divides the hub nodes into block nodes and boundary nodes, calculates the intermediate parameters of the block nodes and the intermediate parameters of the boundary nodes respectively, and can split the original single-layer intermediate parameters into multiple layers, which can greatly compress the magnitude of the intermediate parameters. Taking the expressway road network of a certain province at the beginning of 2024 as an example, it has 103 hub nodes, and these 103 hub nodes contain 218 virtual station codes, and more than 47,000 (218 * 218) intermediate parameters need to be calculated; after dividing the 103 hub nodes into 5 areas, on average, each partition needs to calculate the intermediate parameters of about 25 hub nodes (i.e., block nodes) (the boundary nodes belong to multiple partitions, and there are about 50 virtual stations within the partition), and 16 boundary nodes (36 virtual stations). The total number of intermediate data that needs to be calculated is reduced to about 14,000, saving about 70% of the memory space occupation. Description of the Drawings

[0057] Figure 1 It is a flow chart of a calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology of the present invention.

[0058] Figure 2 It is a schematic diagram of the routing node of the present invention.

[0059] Figure 3 It is the road network graph adopted by the present invention.

[0060] Figure 4 It is a schematic diagram after partitioning based on the hub node of the present invention.

[0061] Figure 5 It is a schematic diagram of the boundary hub distribution after partitioning based on the hub node of the present invention.

[0062] Figure 6 It is a schematic diagram of the fusion of three boundary hubs of the present invention.

[0063] Figure 7 It is a schematic diagram of the demarcation line after the fusion of the boundary hubs of the present invention.

[0064] Figure 8 It is a schematic diagram of the fusion of two boundary hubs of the present invention.

[0065] Figure 9 It is a schematic diagram of the partition after fusing the boundary nodes of the present invention.

[0066] Figure 10 It is a schematic diagram of using intermediate parameters for path fitting of the present invention. Detailed implementation manners

[0067] The present invention will be described below with reference to the accompanying drawings.

[0068] The present invention relates to a calculation method for the minimum toll path based on the graph automatic partitioning technology. The main objective of this method is to solve the problem of excessive memory occupation of the intermediate parameters in the path fitting algorithm, and balance the resource occupation of space and time of the path fitting algorithm by compressing the number of intermediate parameters in the path fitting.

[0069] The processing steps of this method are as Figure 1 shown, and its processing steps include:

[0070] S1, constructing a highway road network model based on charging parameters: constructing a highway road network model based on the routing node information and the mileage amount table information within the section in the charging parameters;

[0071] S2, determining the number of partitions in combination with the actual physical space: determining the number of partitions based on the highway road network model and the number of each traffic hub node within the set geographical space;

[0072] S3. Use the graph automatic partitioning technology to partition the road network: Adopt the K-Means++ algorithm to automatically divide the hub nodes of the entire road network into K regions, record the region labels for each hub node, and use the silhouette coefficient to evaluate the partitioning quality;

[0073] S4. Fuse the partitions to obtain boundary nodes: Generate an alternative boundary node set and a boundary line set based on the hub nodes obtained after partitioning in S3. Select boundary nodes based on the number of times the boundary nodes in the alternative boundary node set appear in the boundary line set and record the partitions to which they belong;

[0074] S5. Calculate the intermediate parameters of the block nodes: Treat any two hub nodes within the same region as block nodes, construct a hub set for the intermediate parameters of the block nodes, and calculate the intermediate parameters of the minimum cost path based on the intermediate parameter amounts in the mileage amount table within the road section as the intermediate parameters of the block nodes;

[0075] S6. Calculate the intermediate parameters of the boundary nodes: Combine the boundary nodes, the intermediate parameters of the block nodes, and the mileage amount table within the road section to calculate the intermediate parameters of the minimum cost path as the intermediate parameters of the boundary nodes;

[0076] S7. Use the intermediate parameters for fast path fitting: Combine and match through the parameters in the mileage amount table within the road section, the intermediate parameters of the block nodes, and the intermediate parameters of the boundary nodes to obtain the minimum cost path between any two toll stations in the road network.

[0077] (1) Construct a highway road network model based on billing parameters

[0078] To partition the road network graph, it is first necessary to construct a road network graph model. The highway road network model can be obtained by processing the routing nodes (hub interchange node information) and the mileage amount table information in the billing parameters.

[0079] The routing node represents the connection relationship between road sections in the highway road network model. Routing node information will be generated at the intersection of multiple different highways. A routing node records the virtual station codes passing through two different road sections, indicating that one road section has a connection relationship with another road section. For example Figure 2As shown in the figure, Longqiaoluo connects three highway sections, namely the highway section where Longqiao (No. 3805) and Mawu Toll Station (No. 3804) are located, the highway section where Fuling South (No. 3202) Toll Station is located, and the highway section where Fuling West (No. 3108) Toll Station is located. At Longqiaoluo, three different virtual station codes (3898, 3201, 3198) are defined for the three sections respectively, and it will generate three routing node records: 3198 - 3898, 3198 - 3201, and 3201 - 3898; while Maanluo indicates the connection of two sections, and it will generate one routing node record: 3398 - 3899. The mileage and amount table within a section records the mileage, amount, combination of toll units, and combination of ETC gantries between any stations (including virtual stations) within the same section. These toll units and ETC gantries belong to the same section.

[0080] The steps to construct a highway road network model through routing nodes and the mileage and amount table within a section are as follows:

[0081] ① Extract the routing node data from all billing parameters, aggregate the station codes representing the same routing node together, and take the smallest toll station code as the new hub node code to obtain the set N of all hub nodes in the road network. Store the multiple station codes of the same hub node as the attributes of the hub node.

[0082] ② Poll the set N of hub nodes. For each hub node, obtain the data in the mileage and amount table within the section with the station code of the hub node as the starting point and the station code of the hub node as the ending point. For each mileage and amount table within the section, group the data with the first gantry as the key, and retain the data with the smallest path amount in each group to obtain the set N1 of data in the mileage and amount table within the section between the hub node and other adjacent hub nodes.

[0083] ③ Use the hub node as a vertex of the graph and each piece of data in the set N1 as an edge of the graph to construct the road network model G.

[0084] (2) Combine with the actual geographical space to determine the number of partitions in the road network

[0085] Combine with the traffic road network model, analyze the distribution characteristics of key nodes, and manually determine how many regions the entire traffic road network model is divided into, that is, determine the value of K.

[0086] Taking the road network map of a certain province ( Figure 3 ) as an example, after manual analysis, it is planned to be divided into 5 regions, that is, the value of K is 5.

[0087] (3) Use the graph automatic partitioning technology to partition the road network

[0088] Use the K-Means++ algorithm to automatically partition the road network. The initialization method of the K-Means++ algorithm is used to select the initial regional centers. First, randomly select a hub node as the first regional center; for the remaining hub nodes, the algorithm calculates the distance between each point and the selected regional center points, and selects the point with the farthest distance as the new regional center point until K regional center points are selected. Assign each hub node to the nearest regional center point to form K clusters.

[0089] Use the silhouette coefficient to evaluate the quality of the K clusters. The silhouette coefficient is an index to measure the goodness of clustering effect, and its value ranges from -1 to 1. The silhouette coefficient combines the cohesion and separation of the clusters, and is used to evaluate whether the samples in the same cluster are closely connected and the separation degree between different clusters.

[0090] 1) Calculate the silhouette coefficient:

[0091] ① For each hub node, calculate the average distance (denoted as a) between it and all other points in the same cluster.

[0092] ② For each hub node, calculate the minimum average distance (i.e., the average distance from this point to all points in the nearest cluster, denoted as b) between it and the points in the nearest cluster other than its own cluster.

[0093] ③ Calculate the silhouette coefficient of each hub node according to the formula: (b - a) / max(a, b).

[0094] ④ Calculate the average value of the silhouette coefficients of all hub nodes, which is the silhouette coefficient of the entire cluster.

[0095] 2) Evaluate the partition quality:

[0096] If the silhouette coefficient is greater than 0.5, it indicates that the partition effect is better, the similarity of hub nodes within the cluster is high, and the similarity of hub nodes between clusters is low, and the automatic partition process can be ended.

[0097] If the silhouette coefficient is less than or equal to 0.5, re-run the K-Means++ algorithm to initialize different regional centers until the silhouette coefficient is greater than 0.5 after re-evaluation using the silhouette coefficient.

[0098] After the regional logical partition by hub nodes is as Figure 4 shown.

[0099] (IV) Fuse boundary nodes

[0100] After dividing different hub nodes into different regions according to the load, the preliminary division of the road network map has been completed ( Figure 4 ), but it has not been clearly defined which region the toll stations between adjacent regions belong to, such asFigure 5 As shown, after dividing the hub nodes, the adjacent hub nodes in different regions form pairs of boundary hub node combinations, and the toll stations on the middle sections between them do not belong to any partition. To solve this problem, it is necessary to fuse the boundary nodes of each region. Each partition is bounded by the fused boundary nodes, and each toll station has a clear belonging area. The steps to fuse the boundary nodes are as follows:

[0101] ① For each partition, traverse the hub nodes within the partition. If the neighbor nodes of a certain hub node are not all within the partition, then record the connection line formed by this node and the neighbor nodes not within the partition as a boundary connection line in the candidate boundary node set.

[0102] ② Find all the edges between the boundary nodes and the boundary nodes of other regions to obtain the boundary line set.

[0103] ③ Traverse each boundary connection line. For the two hub nodes of a boundary line, select the one that appears more frequently in the boundary connection line set as the boundary node.

[0104] ④ If the occurrence times of the hub nodes of a boundary line are the same, then select the one with the smaller node code as the boundary node.

[0105] ⑤ Record the original partition and associated partition information of this boundary node.

[0106] As Figure 6 shown, the three boundary hubs ABC of the three partitions in the figure form two boundary connection lines A - B and A - C. Since the hub node A appears twice, hub A is defined as the boundary node. It is the common boundary node of Region 1, Region 2, and Region 3. After such a definition, the new partition logical demarcation line will be as Figure 7 shown. The toll station in the middle of A - B belongs to Region 1, and the toll station in the middle of A - C belongs to Region 3.

[0107] As Figure 8 shown, two adjacent partitions in the figure form a boundary connection line A - B. Since the code of hub A is smaller than that of hub B, A is defined as the boundary node.

[0108] The logical partition after the fusion process will be as Figure 9 shown.

[0109] (V) Calculate the intermediate parameters of the minimum toll path between any two hub nodes within the region

[0110] After partitioning the hub nodes and processing the boundary nodes, the intermediate parameters of the minimum toll path between any two hub nodes within the region can be calculated based on the mileage and amount table within the section. The steps are as follows:

[0111] ① For each partition, use the hub node and its associated boundary hub nodes together as the hub set S for calculating intermediate parameters.

[0112] ② Take the Cartesian product of the nodes in the hub set S, initialize the intermediate parameters between any two hub nodes within the region. If the two combined hub nodes are the same, set the amount to 0; otherwise, set it to infinity, and store it in the first queue.

[0113] ③ Put all the intermediate parameters with an amount of 0 into the queue Q1.

[0114] ④ Poll the queue Q1, obtain an intermediate parameter from the head of the queue. Assume the start and end points of this intermediate parameter are a and b. Obtain the data with b as the start point and other adjacent hub nodes within the region as the end points from the mileage amount table within the section, and get the mileage amount table set L1 within the section.

[0115] ⑤ Traverse L1. For a certain piece of data in the mileage amount table within the section, assume the start and end points of this data are b and c respectively. If the amount of the intermediate parameter from a to c is greater than the sum of the amount of the intermediate parameter from a to b and the amount from b to c in the mileage amount table within the section, then update the amount of the intermediate parameter from a to c, and add the intermediate parameter from a to c to the first queue.

[0116] ⑥ Repeat the above process until the polling of the Q1 queue is completed.

[0117] ⑦ In the first queue, for the intermediate parameters of the minimum fare path within the region, filter out the data with an amount of 0 and an amount of infinity, and obtain the intermediate parameters of the final minimum fare path.

[0118] (6) Calculate the intermediate parameters of the minimum fare path between any two boundary nodes

[0119] Combine the intermediate parameters of the minimum fare path between any two hub nodes within the region and the mileage amount table within the section to calculate the intermediate parameters of the minimum fare path between any two boundary nodes. The steps are as follows:

[0120] ① Take the Cartesian product of all the boundary nodes, initialize the intermediate parameters between any two boundary nodes, store them in the second queue. If the start and end points are the same hub, set the amount to 0; otherwise, set it to infinity. For any intermediate parameter m of the boundary nodes, if a record n with the same start and end points can be found in the intermediate parameters within the region or the mileage amount table parameters within the section, then compare the amounts of the two records. If the amount of record n is smaller, then update the amount of m to the amount information of n.

[0121] ② Put all the data with an amount not equal to infinity into the queue Q2

[0122] ③Poll the queue Q2. For a certain intermediate parameter of a boundary node, assume the starting point and ending point of the intermediate parameter are a and b. Obtain the data with b as the starting point and other boundary nodes as the ending points from the mileage amount table within the section and the intermediate parameters within the region, and obtain the data set L2.

[0123] ④Traverse L2 and record the data for a certain parameter. Assume the starting point and ending point of this data are b and c respectively. If the intermediate parameter amount from a to c is greater than the sum of the amount from a to b and the amount from b to c in the parameter record, then update the intermediate parameter amount from a to c and add the intermediate parameter from a to c to the second queue.

[0124] ⑤Loop the above steps ③ and ④ until the polling of Q2 is completed.

[0125] ⑥In the second queue, for the intermediate parameters of the minimum fare path of the boundary nodes, filter out the data with an amount of 0 and an infinite amount, and obtain the final data set of the intermediate parameters between any two boundary nodes.

[0126] (VII) Use the regional intermediate parameters and boundary intermediate parameters for path fitting

[0127] After the above processing, the intermediate parameters we obtained are the mileage amount table parameters within the section, the block node intermediate parameters between any two hub nodes within the region, and the boundary node intermediate parameters between any two boundary nodes, a total of three layers of intermediate parameters. Using these three layers of intermediate parameters, according to Figure 10 the combination method shown, the minimum fare path between any two toll stations can be quickly combined and matched. The steps are as follows:

[0128] ①Set the starting toll station as S1 and the ending toll station as S2.

[0129] ②Through the mileage amount table within the section, find the nearest hub nodes A1, A2 and B1, B2 of S1 and S2 (since a toll station belongs to a single section, at most two such hubs can be found). Calculate the path information from a certain toll station to its nearest hub node.

[0130] ③Determine the regions to which S1 and S2 belong through the adjacent hub nodes of S1 and S2.

[0131] ④Find the minimum fare paths from A1 and A2 to all boundary nodes R1, R2...Rn through the intermediate parameters of the block nodes within the region of the starting toll station; find the minimum fare paths from the boundary nodes R1`, R2`...Rn` to the adjacent hubs B1, B2 of S2 through the intermediate parameters within the region of the ending toll station.

[0132] ⑤Find the minimum fare paths from the starting boundary nodes R1, R2...Rn to the ending boundary nodes R1`, R2`...Rn` through the boundary node intermediate parameters.

[0133] ⑥ Combine the intermediate parameters of steps ②, ④, and ⑤ in the order shown as Figure 10 shown to obtain the set of all alternative path results.

[0134] Among them, search for the minimum toll path information from the starting toll station to the starting hub and from the ending hub to the ending toll station in the mileage amount table within the said section; search for the minimum toll path information from the starting hub to the inner boundary hub of the starting area and from the inner boundary hub of the ending area to the ending hub in the intermediate parameters of the block nodes; search for the minimum toll path information from the inner boundary hub of the starting area to the inner boundary hub of the ending area in the intermediate parameters of the boundary nodes;

[0135] Select the path with the minimum amount in the set as the final result.

[0136] Finally, it should be noted that the above is only used to illustrate the technical solution of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred arrangement scheme, those of ordinary skill in the art should understand that the technical solution of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for calculating the minimum toll path of a highway based on graph automatic partitioning technology, characterized in that: S1, constructing a highway network model based on charging parameters: constructing a highway network model based on routing node information and mileage amount table information in the road section in the charging parameters; S2, determining the number of partitions in combination with the actual physical space: determining the number of partitions based on the highway network model and the number of transportation hub nodes in the set geographical space; S3, using graph automatic partitioning technology to partition the road network: using the K-Means++ algorithm, the hub nodes of the entire road network are automatically divided into K regions and the region label is recorded for each hub node, and the silhouette coefficient is used to evaluate the partition quality; S4, merge partitions to obtain boundary nodes: generate a candidate boundary node set and a boundary line set based on the hub node obtained after S3 partitioning, select boundary nodes and record the partitions to which they belong based on the number of times the boundary nodes of the candidate boundary node set appear in the boundary line set; S5, calculating the intermediate parameters of the block nodes: any two hub nodes in the same area are regarded as block nodes, a hub set of the intermediate parameters of the block nodes is constructed, and the intermediate parameters of the minimum fee path are calculated based on the intermediate parameter amounts in the mileage amount table in the road section as the intermediate parameters of the block nodes; S6, calculating the boundary node intermediate parameters: combining the boundary node, the block node intermediate parameters and the mileage amount table in the road section, calculating the intermediate parameters of the minimum fee path as the boundary node intermediate parameters; S7, use intermediate parameters for fast path fitting: combine and match the mileage amount table parameters in the road section, the block node intermediate parameters and the boundary node intermediate parameters to obtain the minimum fee path between any two toll stations in the road network.

2. The calculation method of the minimum toll path for an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that The method for constructing a highway network model based on charging parameters in S1 is: S11, aggregate routing nodes into hub nodes: extract routing node data from all charging parameters, aggregate station codes representing the same routing node together, take the smallest toll station code as the new hub node code, obtain the set N of all hub nodes in the road network, store multiple station codes of the same hub node as attributes of the hub node, and use them as point information in the road network model; S12, obtaining edge information between hub nodes through the mileage and money table within the road section: the mileage and money table within the road section contains mileage information and money information between any two stations within the same road section; If two hub nodes are adjacent, the mileage and fee information are directly queried through the mileage and fee table within the road section as edge information; if there are multiple hub nodes in the same road section, the mileage and fee table within the road section will have records that cross hub nodes. In order to obtain the connection relationship between adjacent hub nodes, for the mileage and fee table records within the road section with the same starting node, the mileage and fee table records within the road section with the smallest charging unit are taken to generate the connection relationship between the hub nodes as edge information; S13, constructing a highway network model based on the point information and edge information.

3. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that, The method for determining the number of partitions in S2 in combination with the actual physical space is as follows: Based on the threshold of the distribution quantity of the hub nodes in the set geographical space, determine the number of partitions of the entire road network map, and set the number of partitions as the K value.

4. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that The method for partitioning the road network using the graph automatic partitioning technology in S3 is as follows: S31: Use the K-Means++ algorithm to automatically partition the road network: Randomly select a hub node as the first regional center; for the remaining hub nodes, the algorithm calculates the distance between each point and the selected regional center point, and selects the point with the farthest distance as the new regional center point until K regional center points are selected; assign each hub node to the nearest regional center point to form K clusters; S32: Use the silhouette coefficient to evaluate the quality of the K clusters: If the silhouette coefficient is less than or equal to 0.5, re-run the K-Means++ algorithm for re-partitioning until the silhouette coefficient is greater than 0.5 and stop re-partitioning.

5. The calculation method of the minimum toll path for expressways based on the graph automatic partitioning technology according to claim 4, characterized in that The method for calculating the silhouette coefficient is as follows: S3A: For each hub node, calculate the average distance between it and all other points in the same cluster and record it as a; S3B: For each hub node, calculate the minimum average distance between it and the nearest cluster outside its cluster, that is, the average distance from this point to all points in the nearest cluster, and record it as b; S3C: Calculate the silhouette coefficient c of each hub node according to the formula i : S3D: Calculate the average value of the silhouette coefficients of all hub nodes, which is the silhouette coefficient of the entire cluster.

6. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that, The method for fusing the partition boundary nodes in step S4 is as follows: S41: For each partition, traverse the hub nodes in the partition. If the neighbor nodes of a certain hub node are not all within the partition, record the connection line formed by this node and the neighbor nodes not within the partition as the boundary connection line in the alternative boundary node set; S42: Find the edges between all boundary nodes and the boundary nodes of other regions to obtain the boundary line set; S43: Select the one with the most occurrences in the boundary line set as the regional boundary point in the alternative boundary node set. If the number of occurrences is the same, select the one with the smaller node code as the boundary node, and record the original partition and associated partition information of this boundary node.

7. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that, The method for finding the minimum fare path in step S5 is as follows: S51: Regard any hub node within the same region as a block node. If any two of the block nodes belong to the same road section, obtain the minimum fare path through the mileage and fare table within the road section; If there are multiple road sections, select the one with the smallest amount in the amount information as the minimum fare path; S52: If a block node is connected to multiple road sections, regard the block node as a virtual station and find the minimum fare path of the hub node between two connected road sections based on the virtual station code; S53: Repeat steps S51 - S52 until the minimum fare path between any two block nodes is found.

8. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that, The method for calculating the intermediate parameters of the block nodes in step S5 is as follows: S5A: Construct the hub set of intermediate parameters: Regard any hub node within the same region as a block node. For each partition, use the block node and the boundary node associated with the block node together as the hub set S for calculating the intermediate parameters; S5B: Initialize the intermediate parameters of any two block nodes: perform the Cartesian product on the nodes in the hub set S, initialize the intermediate parameters of any two block nodes within the same area and set them as the first queue. If the two combined block nodes are the same, set the amount to 0; otherwise, set it to infinity. S5C: Put all intermediate parameters with an amount of 0 into the queue Q1 as the queue to be allocated. S5D: Poll the queue Q1, obtain an intermediate parameter from the head of the queue. Assume the start and end points of this intermediate parameter are a and b. Obtain the data with b as the start point and other adjacent block nodes of b within the area as the end points from the mileage amount table within the section, and obtain the mileage amount table set L1 within the section. S5E: Traverse L1. For the data in a certain mileage amount table within the section, assume the start and end points of the data are b and c respectively. If the intermediate parameter amount from a to c is greater than the intermediate parameter amount of other path schemes, that is, the intermediate parameter amount from a to b plus the amount from b to c in the mileage amount table within the section, then update the intermediate parameter amount from a to c to the intermediate parameter amount of other path schemes, and add the intermediate parameter from a to c to the first queue. S5F: Repeat the above steps S5A - S5E until all intermediate parameters in the Q1 queue are polled. S5G: For the intermediate parameters of the minimum - fee path within the same area in the first queue, filter out the data with an amount of 0 and an amount of infinity to obtain the intermediate parameters of the block nodes of the final minimum - fee path.

9. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technology according to claim 1, characterized in that, S6 The method for calculating the intermediate parameters of the boundary nodes is as follows: S61: Combine the boundary nodes through the Cartesian product and initialize the amount of the path to infinity. S62: Search for the intermediate parameters of the minimum - fee path within the area, obtain the path intermediate parameters between any two boundary nodes within the same area, and update the path information of the boundary nodes. S63: For the data with the same boundary node as the start and end points, if the intermediate parameter amount with the start and end points of the data as the start and end points is less than the sum of the amounts of the data, then update the path information and amount information of the data. S64: Repeat the iteration of step S63 until the intermediate parameters of the minimum - fee paths between all boundary nodes are calculated and used as the intermediate parameters of the boundary nodes.

10. A calculation method for the minimum toll path of an expressway based on the graph automatic partitioning technique according to claim 1, characterized in that S7 The method for performing fast path fitting using the intermediate parameters is as follows: S71: First, through the mileage amount table within the section, find the nearest hub nodes A1, A2 and B1, B2 of the start toll station and the end toll station respectively. S72: Respectively find the boundary node sets R1, R2…Rn of the partition where the start toll station belongs, and the boundary node sets R1`, R2`…Rn` of the partition where the end toll station belongs. S73: Through the hierarchical connection method from the starting toll station to the starting hub, from the starting hub to the inner boundary hub in the starting area, from the inner boundary hub in the starting area to the inner boundary hub in the ending area, from the inner boundary hub in the ending area to the ending hub, and from the ending hub to the ending toll station, perform the Cartesian product on the node sets at each level respectively, search for the minimum toll path information at each level, and finally concatenate them to obtain multiple path information from the starting point to the ending point; Search for the minimum toll path information from the starting toll station to the starting hub and from the ending hub to the ending toll station in the mileage and toll table within the section; Search for the minimum toll path information from the starting hub to the inner boundary hub in the starting area and from the inner boundary hub in the ending area to the ending hub in the intermediate parameters of the block nodes; Search for the minimum toll path information from the inner boundary hub in the starting area to the inner boundary hub in the ending area in the intermediate parameters of the boundary nodes; S74: For multiple toll path information, select the toll path information with the minimum amount as the minimum toll path between any two toll stations in the road network.

Images