Public transport route generation method and system based on road congestion avoidance

Abstract

The application discloses a public transport line generation method and system based on road congestion avoidance, comprising the following steps: constructing a public transport search graph; obtaining a pickup point input by a user, obtaining a congestion index of the input pickup point, a first-order neighbor of the input pickup point or a second-order neighbor of the input pickup point based on the public transport search graph, and generating a recommended pickup point according to the congestion index; obtaining a starting point and a terminal point input by the user, constructing a candidate line set based on the public transport search graph, screening the candidate line set, and generating a recommended line. The application can provide a road congestion avoidance travel line for the user, save the travel time of the user, intelligently guide the user to go to a suitable station to take a bus during a peak period, enhance the willingness of the user to use public transport for travel, and realize green travel.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of public transportation, in particular to a public transportation route generation method and system based on road congestion avoidance. BACKGROUND

[0002] With the rapid development of China's economy, the number of motor vehicles is growing rapidly, which has brought great challenges to urban traffic. The widespread traffic congestion problem seriously affects people's travel experience. In order to improve the urban road congestion situation, in June 2019, the Ministry of Transport and other twelve departments jointly issued the "Green Travel Action Plan (2019-2022)", encouraging citizens to use public transportation.

[0003] When taking public transportation, the accessibility of the boarding route greatly affects the willingness of citizens to choose, which mainly involves factors such as travel time, waiting time, transfer times, walking distance, road congestion, etc. When there are multiple travel routes within a certain range around the departure point to reach the destination, how to comprehensively consider the above factors to select the boarding station and travel route becomes a problem of practical significance.

[0004] Some current commercial software provides a solution to the above problem, which can give multiple alternative routes for public transportation according to the user's travel start and end point information. However, many solutions are still based on the static search results provided by the urban public transportation network, without considering the dynamic road conditions during actual travel. Moreover, almost all software adopts the nearest boarding strategy without considering the guidance of the user's boarding station, which leads to the difficulty of making appropriate travel plans for users when the road traffic pressure is high, and seriously affects the willingness of users to choose public transportation. SUMMARY

[0005] The main purpose of the present application is to provide a public transportation route generation method and system based on road congestion avoidance, which guides the departure from the boarding station, perceives the real-time congestion of the roads around the departure station, and guides the user to the best subway station or bus station to take the bus. Then, based on the determination of the boarding station, the real-time congestion of the roads along the starting point and the end point is considered, and the K-shortest path and the depth-first search method are combined to use the public transportation route search algorithm based on road congestion avoidance to provide the user with a smooth and convenient public transportation route.

[0006] The present application adopts the following technical solutions:

[0007] On the one hand, a public transportation route generation method based on road congestion avoidance includes the following steps:

[0008] Construct a public transportation search graph;

[0009] Obtain the pick-up point input by the user; based on the public transportation search map, obtain the congestion index of the input pick-up point and its first-order or second-order neighbors; and generate a recommended pick-up point based on the congestion index.

[0010] Obtain the user's input start and end points, construct a candidate route set based on the public transportation search map, filter the candidate route set, and generate recommended routes.

[0011] Preferably, a public transportation search map is constructed, specifically including:

[0012] (11) A certain station a of a certain public transportation mode i Construct a circle with center r and radius r. i The circular area of ​​influence, and the stations of different public transportation modes within that area. j That is, with a i These are adjacent transfer stations, at a i With b j A two-way transfer side is added between them; the public transportation modes include subway and bus;

[0013] (12) Let the edge weight be the congestion coefficient between adjacent stations. The congestion coefficient is calculated as follows:

[0014]

[0015] Where, δ i,j C represents the congestion coefficient of road segment i in the j-th time period; i,j R represents the average speed of road segment i during the j-th time period; i,j This represents the reference speed value of road segment i during the j-th time period;

[0016] (13) Based on the transfer edge and the congestion coefficient, the directed graph G is defined as follows:

[0017]

[0018] Among them, G bus Represents a city bus network map; G URT This represents a rail transit network diagram; V represents the set of vertices of G; V bus G represents bus The set of vertices; V URT G represents URT The set of vertices of G; E represents the set of edges of G; E bus G represents bus The set of edges; E URT G URT The set of edges; E transferrepresents a transfer edge set; any vertex v e V in G represents a public transportation station, including a set of public transportation lines and attribute information of longitude and latitude position; any directed edge e = (v i , v j ) e E in G represents a public transportation line from vertex v i to vertex v j , including a set of public transportation lines and congestion coefficient attribute information; the congestion coefficient is used as the weight of each directed edge on the way, and the congestion coefficients of the directed edges e e E URT U E transfer are set to 0.

[0019] Preferably, the input pickup point of the user is obtained, the congestion index of the input pickup point, the first-order neighbor or the second-order neighbor of the input pickup point is obtained based on the public transportation search graph, and the recommended pickup point is generated according to the congestion index, specifically including:

[0020] (21) After obtaining the input pickup station of the user, the first-order neighbor of the pickup station is searched according to the constructed public transportation search graph; the first-order neighbor is a directly reachable node of the pickup station;

[0021] (22) The congestion index of the input pickup station of the user to all first-order neighbors thereof is calculated, as follows:

[0022]

[0023] wherein r p,q represents the congestion index of the connection section from the input pickup station v p to the other reachable station v q ; s t_p,q represents the real-time vehicle speed of the connection section from the input pickup station v p to the other reachable station v q ; s avg_p,q represents the historical average vehicle speed of the connection section from the input pickup station v p to the other reachable station v q , respectively.

[0024] (23) If the congestion index exceeds the threshold value η, it is judged that the connection section is congested, and the number of congested connection sections is calculated;

[0025] (24) If the number of congested connection sections exceeds half of the first-order neighbors, a new pickup station is searched, and (25) is performed; otherwise, the input pickup station of the user is taken as the pickup station;

[0026] (25) If there is a subway station near the input pickup station of the user, subway is preferentially recommended;

[0027] (26) If there is no subway station near the user-inputted pick-up station, check whether there is a first-order neighbor station with a connection road segment congestion index less than a preset value η, select the one with the smallest congestion index as the new pick-up station; if all first-order neighbor stations have connection road segments greater than η, search for second-order neighbor stations, repeat step (21); the second-order neighbor is a direct node of the first-order neighbor of the pick-up station;

[0028] (27) If the connection road segments of the second-order neighbor stations are all greater than η, select the user-inputted pick-up station for the nearest ride.

[0029] Preferably, the start point and the end point inputted by the user are acquired, a candidate route set is constructed based on the public transport search graph, the candidate route set is screened, and a recommended route is generated, specifically comprising:

[0030] (31) After the user inputs the start point and the end point, a direct route between the start point and the end point is searched;

[0031] (32) A KSP algorithm is used to search for a plurality of base paths with a congestion coefficient lower than a preset congestion value;

[0032] (33) On the base path, a public transport travel route meeting the requirements is acquired according to actual needs, as follows:

[0033] (34) All public transport lines of the start point station are traversed, and the longest co-line distance thereof from the base path is calculated;

[0034] (35) If the longest co-line distance is the base path, the candidate route set is added;

[0035] (36) Otherwise, the base path is truncated based on the subscript of the longest co-line station, and steps (34)-(36) are recursively executed, and the number of recursive layers does not exceed the number of transfers;

[0036] (37) The candidate route set is screened and sorted;

[0037] (38) The final recommended route is outputted.

[0038] On the other hand, a public transport route generation system based on road congestion avoidance comprises:

[0039] A search graph construction module is configured to construct a public transport search graph;

[0040] A pick-up point recommendation module is configured to acquire a user-inputted pick-up point, and to acquire, based on the public transport search graph, a congestion index of the input pick-up point, a first-order neighbor or a second-order neighbor of the input pick-up point, and to generate a recommended pick-up point according to the congestion index;

[0041] A route recommendation module is configured to acquire a start point and an end point input by a user, construct a candidate route set based on the public transport search graph, filter the candidate route set, and generate a recommended route.

[0042] From the above description of the present application, the present application has the following beneficial effects:

[0043] The present application provides a public transport route generation method and system based on road congestion avoidance. The method guides a user to the best subway station or bus station for boarding from a boarding station. Then, based on the boarding station, the real-time congestion status of the roads along the start point and end point is considered, and a public transport route search algorithm based on road congestion avoidance is used to provide a smooth and convenient public transport route for the user, thereby saving the user's travel time, especially during peak hours, intelligently guiding the user to the appropriate station for boarding, enhancing the user's willingness to use public transport, and achieving green travel. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 A flowchart of the public transport route generation method based on road congestion avoidance of the present application;

[0045] Figure 2 A public transport route search graph structure used in the present application;

[0046] Figure 3 A k-order neighbor diagram;

[0047] Figure 4 A RCA-GS flowchart based on road conditions in the present application;

[0048] Figure 5 A RCA-RS flowchart based on congestion avoidance in the present application;

[0049] Figure 6 A comparison chart of the optimal scheme recommended by the RCA-RS and the travel time under ideal conditions;

[0050] Figure 7 A structural block diagram of the public transport route generation system based on road congestion avoidance of the present application. DETAILED DESCRIPTION

[0051] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0052] like Figure 1 As shown, the present invention provides a method for generating public transportation routes based on road congestion avoidance, comprising the following steps:

[0053] S101, construct a public transportation search map;

[0054] S102, obtain the pick-up point input by the user, and based on the public transportation search map, obtain the congestion index of the input pick-up point and the first-order or second-order neighbors of the input pick-up point, and generate a recommended pick-up point according to the congestion index.

[0055] S103: Obtain the user's input start and end points, construct a candidate route set based on the public transportation search map, filter the candidate route set, and generate recommended routes.

[0056] Specifically, the following will provide a detailed explanation of how to construct a public transportation search map.

[0057] This example uses the public transportation operation map of Siming District, Xiamen City, Fujian Province in April 2021. The dataset contains 297 bus routes and 2 subway lines, and the speed data is from 17:11 on June 3, 2021.

[0058] like Figure 2 As shown, in urban public transportation networks, buses and rail transit have their own independent network diagrams. To generate a complete and unified method for public transportation travel routes, they need to be merged. This requires adding transfer edges between bus stops and rail stops to the network diagram. Assuming a user transfers between different public transportation modes on foot, then a stop a for a certain public transportation mode... i Construct a circle with center r and radius r. i The circular area of ​​influence, with different public transportation stops within that area. j That is, with a i These are adjacent transfer stations, at a i With b j Add a bidirectional transfer edge between them. Here, r is set to 500 meters.

[0059] In the search graph structure, the edge weight is the congestion coefficient between adjacent stations. In this construction method, the INRIX congestion coefficient is used for calculation, and its calculation formula is shown in Equation 1:

[0060]

[0061] Where, δ i,j C i,j R i,j These represent the congestion coefficient, average speed, and reference speed of road segment i during the j-th time period, respectively.

[0062] Based on the above transfer edges, road congestion coefficients, and the city bus network map G bus Rail transit network map G URT The directed graph G used in this invention is defined as shown in Equation 2:

[0063]

[0064] Among them, V, V bus V URT G and G respectively bus G URT The set of vertices; E, E bus E URT G and G respectively bus G URT The set of edges, E transfer Let G represent the set of transfer edges. Any vertex v∈V in G represents a public transportation station, possessing attributes such as the set of public transportation lines stopping at that station and its latitude and longitude location; any directed edge e = (v... i ,v j )∈E, indicating that there is a public transportation route from vertex v i Driving to the top v j It possesses attributes such as a set of public transportation routes and congestion coefficients. In RCA, the congestion coefficient is used as the weight of each directed edge along the route, and the directed edge e∈E is defined as follows. URT ∪E transfer The congestion coefficient is 0.

[0065] Specifically, the steps of the S102 Road Condition-Based Boarding Station Guidance Strategy (RCA-GS) will be explained in detail below.

[0066] To address the problem of users struggling to determine their boarding point during peak traffic hours, especially when road conditions are heavy, this invention proposes a road condition-based boarding station guidance strategy (RCA-GS).

[0067] First, given a vertex v in a directed graph G... i The set of k-th order neighbor vertices V i (k) The definition is shown in Equation 3:

[0068] V i (k) ={v j |v j ∈G,|v i →v j |=k} (9)

[0069] Among them, |v i →v j |=k indicates that there exists a connected path between vertices vi and vj with a shortest distance of k (k≥0). Also, let V... i (0) ={v i A diagram of the k-order neighbors is shown below. Figure 3 As shown, assuming node a is the intended departure station, then nodes b, c, e, and f are its first-order neighbors, and nodes d and g are its second-order neighbors.

[0070] According to the GB-50200-95 standard, the distance between stations on urban ground transportation routes is 500–800 meters. Based on statistical experience, the maximum walking distance that users can generally tolerate is 1000–1500 meters. Therefore, in RCA-GS, k = 2 is chosen. A station v is defined. i The set of reachable sites V i =V i (1) ∪V i (2) .

[0071] The flowchart of RCA-GS is as follows: Figure 4 As shown:

[0072] After the user inputs their desired boarding station, the congestion index from that station to all its first-order neighbors is calculated using the formula shown in Equation 4.

[0073]

[0074] Where, r p,q This indicates the input boarding station v. p To another reachable site v q The congestion index of the connecting road sections; s t_p,q This indicates the input boarding station v. p To another reachable site v q The real-time vehicle speed of the connecting road segment; s avg_p,q These represent the input boarding station v. p To another reachable site v q The historical average vehicle speed of the connecting road sections.

[0075] According to formula 4, if the congestion index exceeds the threshold η, it is considered that the connection road section is congested, and the number of congested connection road sections is calculated. If more than half of the first-order neighbors of the connection road section are in a congested state, the user is advised to change the boarding station. When changing the boarding station, if it is detected that there is a rail station around the current first-order neighbor, the nearest rail station is set as the new boarding station; otherwise, it is checked whether there is a first-order neighbor station with a connection road section congestion index less than η, and the one with the smallest congestion index is selected as the new boarding station; if all the connection stations of the first-order neighbors are greater than η, the second-order neighbor stations are searched, and the above operation is repeated; if the connection road sections of the second-order neighbor stations are all greater than η, the initial boarding station is selected to take the train nearby (or it is suggested to change to other transportation modes).

[0076] In order to verify the effectiveness of RCA-GS, 30 OD pairs were randomly selected multiple times in pure bus mode and bus-metro alleviating mode, and the average congestion coefficients of all lines before and after guidance were compared, and the results are shown in Table 1. The search graph established in step 1 is used in the experiment.

[0077] Table 1 Congestion coefficients before and after guidance

[0078]

[0079]

[0080] As can be seen from Table 1, the congestion coefficients of the routes before the station guidance are higher than those after the station guidance. For the bus-metro mixed mode, the congestion coefficients after the guidance are reduced by about 12% compared with those before the guidance; for the pure bus mode, they are reduced by about 15%. This shows that the station guidance strategy has a certain effect of avoiding road congestion.

[0081] Specifically, the steps of S103, route search of public transportation based on congestion avoidance (RCA-RS), will be described in detail as follows.

[0082] In RCA-RS, first, a route search sub-algorithm RCA-RS-K is designed based on K shortest path and path truncation, and the pseudo code is shown in Table 2. The KSP algorithm uses Yen algorithm, which can generate k paths based on the structure of directed acyclic graph with non-negative weight edges, without relying on a certain specific shortest path algorithm (Dijkastra algorithm is used in RCA-RS).

[0083] Compared with other algorithms that calculate the congestion of the route after generating the route scheme, RCA-RS-K considers the road congestion during route search and combines the truncation operation, which helps to reduce unnecessary search overhead.

[0084] Table 2 Pseudo code of RCA-RS-K algorithm

[0085]

[0086]

[0087] According to the pseudo code in Table 2, the main flow is described as follows:

[0088] 1) First search the direct public transport routes between the start and end points, and then use Yen algorithm to calculate k relatively uncongested paths as base paths;

[0089] 2) For each base path p, traverse all public transport routes of the boarding station: detect the farthest station p' of each public transport route that is co-linear with p. If p' is the end station, a candidate route scheme is generated; otherwise, take p' as a new boarding station, truncate the path p to obtain its subsequent part as a new base path, and recursively execute the above process. The recursion level does not exceed the number of transfers m.

[0090] 3) After the base path search is completed, the candidate route scheme set is processed through a filter filter: scheme screening and scheme index calculation.

[0091] In scheme screening, the following is mainly considered to remove unreasonable schemes: (1) whether there is a same station transfer; (2) if it cannot reach the end point directly, whether it can reach the end point by walking within an acceptable distance; (3) whether there is an unreasonable walking mode.

[0092] The index calculation mainly includes three indexes: congestion factor δ, travel time T, and number of transfers m, and the calculation formulas are shown in formulas 5 and 6.

[0093]

[0094]

[0095] In formula 5, N all represents the number of public transport station pairs passed by the route scheme, N con represents the number of station pairs where traffic congestion occurs, and the judgment of congestion is based on formula 4. In formula 6, dh represents the spherical distance between two points, u represents the non-straight-line coefficient of urban bus roads, v1 represents the real-time speed or predicted speed of the bus, v2 represents the stable running speed of non-bus modes (such as walking, rail), and Twait represents the waiting time.

[0096] 1) For the screened travel route scheme set, sort it in the order of considering transfer, congestion, and then time.

[0097] The RCA-RS-K sub-algorithm can quickly search and generate a set of travel route options, but there may be situations where no effective solutions can be found for certain origin and destination points. This is because congestion avoidance needs to be prioritized, which may result in some travel routes involving multiple transfers, making such routes unsuitable for users. To address this issue and enhance the robustness of the RCA-RS algorithm, a depth-first search (DFS) method is introduced, and a travel route search sub-algorithm, RCA-RS-D, is designed. The pseudocode is shown in Table 3.

[0098] RCA-RS-D will be triggered when RCA-RS-K fails to output a valid travel route. Its main process is as follows: iterate through all bus routes at the boarding station and select routes by default according to the RCA-RS-K method; otherwise, use the RCA-RS-D algorithm to perform a recursive search, with the recursion level not exceeding a reasonable number of transfers.

[0099] Table 3. Pseudocode of the RCA-RD-D algorithm

[0100]

[0101]

[0102] To verify the capabilities of the RCA-RS algorithm, the following experiments were conducted: a comparison experiment of travel time, a comparison experiment of the number of effective routes, and an experiment on the effectiveness of the solutions provided by the algorithm. The search graph established in step 1 was used in the experiments.

[0103] • Travel time comparison experiment

[0104] One hundred OD pairs were randomly selected, and the ideal travel time for each OD pair was compared with the travel time of the optimal route obtained by the RCA-RS algorithm. Here, the ideal travel time refers to the optimal route travel time obtained by simulating people's travel process using only the traditional DFS algorithm, without considering road conditions. Experimental results are as follows: Figure 6 As shown.

[0105] like Figure 6 As shown, the optimal route travel time obtained by RCA-RS is generally longer than the ideal time, which is due to the consideration of road traffic conditions; however, there are some cases where the optimal route time of RCA-RS is shorter than the ideal time. This is because the ideal case uses the historical average speed of the corresponding road surface for calculation, while the road surface conditions change in real time, and there are cases where the actual speed is faster than the average speed.

[0106] It can be found from the figure that the two curves have good fitting degree, and further calculation and analysis can obtain the average time difference of 1.8 minutes, which shows that the RCA-RS mechanism can provide the travel route close to the ideal travel time for the user to select.

[0107] · Effective line number comparison experiment

[0108] The number of effective travel routes determines the diversity of the user's selection options. In the experiment, the number of travel routes provided by RCA-RS-K (based on KSP), DFS, and RCA-RS-D (based on the above two) is compared. Three modes of pure bus, pure subway, and bus-subway mixed are used, and 100 OD pairs are randomly selected multiple times. The experimental results are shown in Table 4:

[0109] Table 4 Comparison of route numbers provided by different algorithms

[0110]

[0111] From the experimental results, it can be found that the number of routes provided by the DFS algorithm is the most, except for the pure subway mode. This is because DFS does not consider road congestion when searching for travel options, but only considers the connection between routes. Since bus routes are usually more than subway routes, there are more bus travel options and fewer subway options. The RCA-RS-K algorithm based on KSP provides fewer routes, because through the base path truncation method, there are some OD pairs that cannot obtain feasible routes. The algorithm that combines the two (RCA-RS-D) provides more routes than the RCA-RS-K algorithm, with an average of about 3 optimal routes for each OD pair to choose from.

[0112] · Algorithm provides scheme effectiveness experiment

[0113] In order to verify the feasibility of the travel routes obtained by RCA in real life, the travel route provided by the subway is taken as the baseline for comparison test. Similarly, 100 OD pairs are randomly selected multiple times, and the results are obtained using pure bus mode, pure subway mode, and bus-subway mixed mode. The route hit rate is calculated. The experimental results are shown in Table 5.

[0114] Table 5 Route hit rate experimental results

[0115]

[0116] From Table 5, it can be found that the RCA algorithm results have a good hit rate compared to the results provided by the subway. The hit rate of the pure subway mode is the highest, reaching 92%; for other modes, the scheme hit rate is also more than 50%, which shows that the public transportation travel scheme provided by the algorithm has a certain feasibility in life.

[0117] Referring to Figure 7 As shown in the figures, the application also discloses a public transport route generation system based on road congestion avoidance, comprising:

[0118] A search graph construction module 701 is configured to construct a public transport search graph.

[0119] A pickup point recommendation module 702 is configured to obtain a user-input pickup point, obtain congestion indexes of the input pickup point, first-order neighbors or second-order neighbors of the input pickup point based on the public transport search graph, and generate a recommended pickup point according to the congestion indexes.

[0120] A route recommendation module 703 is configured to obtain a user-input starting point and ending point, construct a candidate route set based on the public transport search graph, filter the candidate route set, and generate a recommended route.

[0121] Although the specific embodiments of the application are described above, those skilled in the art should understand that the specific examples described are only illustrative, and are not intended to limit the scope of the application, and equivalent modifications and variations made by those skilled in the art in accordance with the spirit of the application should be covered within the scope of the claims of the application.

Claims

1. A method for generating public transportation routes based on road congestion avoidance, characterized in that, Includes the following steps: Construct a public transportation search map; Obtain the pick-up point input by the user; based on the public transportation search map, obtain the congestion index of the input pick-up point and its first-order or second-order neighbors; and generate a recommended pick-up point based on the congestion index. Obtain the user's input start and end points, construct a candidate route set based on the public transportation search map, filter the candidate route set, and generate recommended routes; Obtain the user-inputted pick-up point. Based on the public transportation search map, obtain the congestion index of the input pick-up point and its first-order or second-order neighbors. Generate recommended pick-up points based on the congestion index, specifically including: (21) After obtaining the boarding station input by the user, search for the first-order neighbors of the boarding station according to the constructed public transportation search graph; the first-order neighbors are the directly reachable nodes of the boarding station; (22) Calculate the congestion index from the user-input boarding station to all its first-order neighbors, as follows: ； Where, r p,q This indicates the input boarding station v. p To another reachable site v q The congestion index of the connecting road sections; s t_p,q This indicates the input boarding station v. p To another reachable site v q The real-time vehicle speed of the connecting road segment; s avg_p,q These represent the input boarding station v. p To another reachable site v q The historical average vehicle speed of the connecting road sections; (23) If the congestion index exceeds the threshold ŋ, the connecting road segment is determined to be congested, and the number of congested connecting road segments is calculated. (24) If the number of congested connecting segments exceeds half of the first-order neighbors, search for a new boarding station and execute (25); otherwise, use the boarding station entered by the user as the boarding station. (25) If there is a subway station near the boarding station entered by the user, the subway will be recommended as the preferred option; (26) If there is no subway station near the boarding station entered by the user, check if there are any first-order neighbor stations with a congestion index of less than the preset value ŋ, and select the one with the smallest congestion index as the new boarding station; if the congestion index of all first-order neighbor stations is greater than ŋ, search for second-order neighbor stations and repeat step (21); the second-order neighbor is the directly reachable node of the first-order neighbor of the boarding station. (27) If the connecting segments of second-order neighboring stations are all greater than ŋ, then select the nearest boarding station entered by the user to take the bus.

2. The public transportation route generation method based on road congestion avoidance according to claim 1, characterized in that, Constructing a public transportation search map specifically includes: (11) A station a of a certain mode of public transportation i Construct a circle with center r and radius r. i The circular area of ​​influence, and the stations of different public transportation modes within that area. j That is, with a i These are adjacent transfer stations, at a i With b j A two-way transfer side is added between them; the public transportation modes include subway and bus; (12) Let the edge weight be the congestion coefficient between adjacent stations. The congestion coefficient is calculated as follows: ； Where, δ i,j C represents the congestion coefficient of road segment i in the j-th time period; i,j R represents the average speed of road segment i during the j-th time period; i,j This represents the reference speed value of road segment i during the j-th time period; (13) Based on the transfer edge and the congestion coefficient, the directed graph G is defined as follows: ； Among them, G bus Represents a city bus network map; G URT This represents a rail transit network diagram; V represents the set of vertices of G; V bus G represents bus The set of vertices; V URT G represents URT The set of vertices of G; E represents the set of edges of G; E bus G represents bus The set of edges; E URT G URT The set of edges; E transfer Represents the set of interchange edges; any vertex in G. , represents a public transportation station, including the set of public transportation lines stopping at the station and its latitude and longitude location attributes; any directed edge in G This indicates that there is a public transportation line from vertex v. i Driving to the top v j This includes attribute information such as the set of public transportation routes and congestion coefficients; the congestion coefficient is used as the weight of each directed edge along the route, and the directed edges are defined. The congestion coefficient is 0.

3. The public transportation route generation method based on road congestion avoidance according to claim 2, characterized in that, Obtain the user's input start and end points, construct a candidate route set based on the public transportation search map, filter the candidate route set, and generate recommended routes, specifically including: (31) After the user enters the starting point and the destination, search for direct routes between the starting point and the destination; (32) Use the KSP algorithm to search for several base paths with congestion coefficients lower than the preset congestion value; (33) Based on actual needs, obtain the required public transportation routes along the base route, as follows: (34) Traverse all public transportation lines at the origin station and calculate the longest collinear distance between it and the base path; (35) If the longest collinear distance is the base path, then add it to the candidate path set; (36) Otherwise, take the index of the longest collinear station as the benchmark, perform a truncation operation on the base path, and recursively execute (34)-(36), with the number of recursion layers not exceeding the number of transfers; (37) Filter and sort the candidate route set; (38) Output the final recommended route.

4. A public transportation route generation system based on road congestion avoidance, characterized in that, include: The search graph building module is used to build public transportation search graphs; The boarding point recommendation module is used to obtain the boarding point input by the user, and to obtain the congestion index of the input boarding point and its first-order or second-order neighbors based on the public transportation search map, and to generate recommended boarding points based on the congestion index. The route recommendation module is used to obtain the user's input start and end points, construct a candidate route set based on the public transportation search map, filter the candidate route set, and generate recommended routes. The boarding point recommendation module is specifically used for: (21) After obtaining the boarding station input by the user, search for the first-order neighbors of the boarding station according to the constructed public transportation search graph; the first-order neighbors are the directly reachable nodes of the boarding station; (22) Calculate the congestion index from the user-input boarding station to all its first-order neighbors, as follows: ； Where, r p,q This indicates the input boarding station v. p To another reachable site v q The congestion index of the connecting road sections; s t_p,q This indicates the input boarding station v. p To another reachable site v q The real-time vehicle speed of the connecting road segment; s avg_p,q These represent the input boarding station v. p To another reachable site v q The historical average vehicle speed of the connecting road sections; (23) If the congestion index exceeds the threshold ŋ, the connecting road segment is determined to be congested, and the number of congested connecting road segments is calculated. (24) If the number of congested connecting segments exceeds half of the first-order neighbors, search for a new boarding station and execute (25); otherwise, use the boarding station entered by the user as the boarding station. (25) If there is a subway station near the boarding station entered by the user, the subway will be recommended as the preferred option; (26) If there is no subway station near the boarding station entered by the user, check if there are any first-order neighbor stations with a congestion index of less than the preset value ŋ, and select the one with the smallest congestion index as the new boarding station; if the congestion index of all first-order neighbor stations is greater than ŋ, search for second-order neighbor stations and repeat step (21); the second-order neighbor is the directly reachable node of the first-order neighbor of the boarding station. (27) If the connecting segments of second-order neighboring stations are all greater than ŋ, then select the nearest boarding station entered by the user to take the bus.

Images