A method and device for matching a ride-hailing carpool, an electronic device and a storage medium

Abstract

The present application relates to a kind of online car-hailing carpooling matching method, device, electronic equipment and storage medium, belong to vehicle carpooling matching technical field, wherein the online car-hailing carpooling matching method includes: based on the start position of multiple groups of OD, the multiple groups of OD are hierarchically clustered, obtain first clustering result;Based on the directed line segment and end position formed by the start point and end point of each group of OD in each clustering cluster in the first clustering result, the first clustering result is optimized, obtain second clustering result;Based on the travel time between the end point of each two groups of OD in each clustering cluster in the second clustering result, the second clustering result is optimized, obtain third clustering result, and the third clustering result is used as carpooling matching result.The present application carries out carpooling matching according to actual OD, guarantees the effect of carpooling matching, improves the experience of user carpooling.

Description

Technical Field

[0001] This invention relates to the field of vehicle ride-sharing matching technology, and in particular to a ride-hailing ride-sharing matching method, device, electronic device, and storage medium. Background Technology

[0002] With the rapid development of modern urban transportation systems and the widespread penetration of ride-hailing services, how to effectively improve ride-sharing matching efficiency and significantly reduce vehicle empty-running rates has become a key issue that the ride-hailing industry urgently needs to address.

[0003] Traditional ride-sharing matching strategies often focus on people in transportation, considering factors such as driver-passenger relationships, and mostly apply existing machine learning models and algorithms. While these algorithms can process data to some extent, they fail to fully integrate the actual geographical and road conditions. Therefore, the matching results generated by these traditional methods are often too theoretical and idealistic, making it difficult to achieve the expected results in real urban traffic environments. This not only affects the travel experience of passengers but also increases the operational burden of ride-hailing services, hindering the sustainable development of the industry. Summary of the Invention

[0004] In view of this, it is necessary to provide a ride-hailing carpooling matching method, device, electronic device and storage medium to solve the problem that the matching results generated by traditional ride-hailing carpooling matching methods are often too theoretical and idealistic, and are difficult to achieve the expected results in the actual urban traffic environment.

[0005] To address the above problems, this invention provides a ride-hailing carpooling matching method, comprising:

[0006] Based on the starting positions of multiple groups of ODs, hierarchical clustering is performed on the multiple groups of ODs to obtain the first clustering result;

[0007] Based on the directed line segments formed by the start and end points of each OD group within each cluster in the first clustering result, and the position of the end point, the first clustering result is optimized to obtain the second clustering result.

[0008] Based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, the second clustering result is optimized to obtain a third clustering result, and the third clustering result is used as the ride-sharing matching result.

[0009] In one possible implementation, the hierarchical clustering of the multiple groups of ODs based on their starting positions to obtain a first clustering result includes:

[0010] Based on the starting positions of the multiple groups of ODs and the Birch algorithm, hierarchical clustering is performed on the multiple groups of ODs to obtain the first clustering result. The threshold parameter and classification factor of the Birch algorithm are determined based on the size of the OD data in different time periods.

[0011] In one possible implementation, the optimization of the first clustering result based on the directed line segments and endpoint positions formed by the start and end points of each group of ODs within each cluster in the first clustering result to obtain a second clustering result includes:

[0012] Based on the first clustering result, the angle between the directed line segments formed by the start and end points of every two groups of OD within each cluster, and the distance between the end points, the first clustering result is optimized to obtain the second clustering result.

[0013] In one possible implementation, the first clustering result is optimized based on the angle between the directed line segments formed by the start and end points of every two groups of ODs within each cluster, and the distance between the end points, to obtain a second clustering result, including:

[0014] Based on the first clustering result, the angle between the directed line segments formed by the start and end points of every two groups of OD within each cluster, and the distance between the end points, the degree of opening and closing between every two groups of OD within each cluster is determined.

[0015] In the first clustering result, ODs within each cluster whose openness to any other OD is greater than a preset threshold are removed to obtain the second clustering result.

[0016] In one possible implementation, the degree of opening between every two groups of ODs within each cluster is calculated based on the following formula:

[0017]

[0018]

[0019] in, Indicates the degree of opening / closing between two sets of OD. and For weight parameters, This represents the angle between the directed line segments formed by the start and end points of two sets of origin (OD). This represents the distance between the endpoints of the two sets of ODs. For the preset distance, The angle between the directed line segment formed by the start and end points of one of the two sets of ODs and the due north direction. It represents the angle between the direction of due north and the direction formed by the starting and ending points of the other OD in two OD groups.

[0020] In one possible implementation, the second clustering result is optimized based on the travel time between the destinations of every two groups of ODs within each cluster to obtain a third clustering result, including:

[0021] Based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, a similarity matrix is ​​constructed for each cluster.

[0022] Based on the similarity matrix corresponding to each cluster and the spectral clustering algorithm, the second clustering result is optimized to obtain the third clustering result.

[0023] In one possible implementation, the elements in the similarity matrix corresponding to each cluster are determined based on the following formula:

[0024]

[0025] in, This represents the similarity between any two groups of ODs within each cluster. This represents the travel time between the destinations of every two OD groups within each cluster. , , , , These are preset parameters. The first time threshold, This is the second time threshold.

[0026] The present invention also provides a ride-hailing matching device, comprising:

[0027] The clustering module is used to perform hierarchical clustering on the multiple groups of ODs based on their starting positions to obtain the first clustering result;

[0028] The first optimization module is used to optimize the first clustering result based on the directed line segment formed by the start and end points of each OD group in each cluster and the position of the end point in the first clustering result, so as to obtain the second clustering result.

[0029] The second optimization module is used to optimize the second clustering result based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, to obtain a third clustering result, and to use the third clustering result as the ride-sharing matching result.

[0030] The present invention also provides an electronic device, including a memory and a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the ride-hailing matching method described above.

[0031] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the ride-hailing matching method described above.

[0032] The beneficial effects of this invention are as follows: The ride-sharing matching method, device, electronic device, and storage medium provided by this invention first perform hierarchical clustering on multiple groups of origin-destination (OD) locations to achieve preliminary clustering. Then, the clustering results are further optimized by using the directed line segments formed by the origin and destination of each OD group, the destination location, and the travel time between the destinations of every two OD groups to improve the ride-sharing matching effect and thus enhance the user's ride-sharing experience. This invention performs ride-sharing matching based on actual OD locations, ensuring the effectiveness of ride-sharing matching and improving the user's ride-sharing experience. Attached Figure Description

[0033] Figure 1 This is a flowchart illustrating an embodiment of the ride-hailing matching method provided by the present invention.

[0034] Figure 2 This is a flowchart illustrating an embodiment of the preliminary clustering process provided by the present invention;

[0035] Figure 3 This is a flowchart illustrating an embodiment of the vector angle calculation process provided by the present invention.

[0036] Figure 4 A flowchart illustrating an embodiment of the opening and closing constraint calculation process provided by the present invention;

[0037] Figure 5 This is a flowchart illustrating an embodiment of the spectral clustering process provided by the present invention;

[0038] Figure 6 A schematic diagram of an embodiment of the ride-hailing matching device provided by the present invention;

[0039] Figure 7 A schematic diagram of the structure of an embodiment of the electronic device provided by the present invention. Detailed Implementation

[0040] Preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings, which form part of this application and are used together with the embodiments of the present invention to illustrate the principles of the present invention, but are not intended to limit the scope of the present invention.

[0041] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. Furthermore, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0042] In the description of this invention, reference to "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the described embodiments can be combined with other embodiments.

[0043] With the rapid development of modern urban transportation systems and the widespread penetration of ride-hailing services, how to effectively improve ride-sharing matching efficiency and significantly reduce vehicle empty-running rates has become a key issue that the ride-hailing industry urgently needs to address.

[0044] Traditional ride-sharing matching strategies often focus on people in transportation, considering factors such as driver-passenger relationships, and mostly apply existing machine learning models and algorithms. While these algorithms can process data to some extent, they fail to fully integrate the actual geographical and road conditions. Therefore, the matching results generated by these traditional methods are often too theoretical and idealistic, making it difficult to achieve the expected results in real urban traffic environments. This not only affects the travel experience of passengers but also increases the operational burden of ride-hailing services, hindering the sustainable development of the industry.

[0045] To address the aforementioned problems, this invention provides a ride-hailing carpooling matching method.

[0046] The specific embodiments are described in detail below:

[0047] A specific embodiment of the present invention discloses a ride-hailing carpooling matching method, combined with Figure 1 Let's take a look. Figure 1 This is a flowchart illustrating an embodiment of the ride-hailing matching method provided by the present invention, including steps S101 to S103, wherein:

[0048] In step S101, based on the starting positions of multiple groups of ODs, hierarchical clustering is performed on the multiple groups of ODs to obtain the first clustering result;

[0049] In step S102, based on the directed line segments formed by the start and end points of each group of ODs within each cluster in the first clustering result, and the position of the end point, the first clustering result is optimized to obtain the second clustering result.

[0050] In step S103, based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, the second clustering result is optimized to obtain a third clustering result, and the third clustering result is used as the ride-sharing matching result.

[0051] In this invention, the origin and destination locations of each Origin-Destination (OD) group can be obtained from the user's order information in the ride-hailing software. After the user initiates an order via a mobile device such as a smartphone, the backend physical server or cloud server can obtain the user's order information through the operator's base station, and then obtain the origin and destination locations of each OD group from the user's order information. Subsequent clustering operations and optimization processes can be implemented through the built-in algorithm in the ride-hailing software. The built-in algorithm can be pre-written and stored on the hard drive of the physical server or cloud server, and then the processor in the physical server or cloud server calls the algorithm to perform data operations, thereby obtaining the ride-sharing matching results. Finally, the physical server or cloud server can send the ride-sharing matching results to the ride-hailing driver's mobile device such as a smartphone through the operator's base station, so that the driver can select a suitable ride-sharing order to accept and provide service.

[0052] In practice, the first step is to perform hierarchical clustering on multiple groups of ODs based on their starting positions to obtain the first clustering result, thereby completing the initial clustering of multiple groups of ODs and assigning ODs with similar starting positions to the same cluster.

[0053] Next, based on the directed line segments formed by the start and end points of each OD group within each cluster, and the end point positions, the first clustering result can be optimized to obtain the second clustering result. The directed line segments formed by the start and end points of each OD group within each cluster can determine the path direction between two OD groups, and the end point positions of each OD group within each cluster can determine the distance between the end point positions of two OD groups, thus allowing for further subdivision of ODs with similar paths.

[0054] Finally, based on the travel time between the destinations of every two OD groups within each cluster, the second clustering result can be optimized to obtain a third clustering result, which is then used as the carpooling matching result. By determining the travel time between the destinations of every two OD groups within each cluster, the path similarity between the two OD groups after carpooling can be further identified, improving the effectiveness of carpooling matching.

[0055] Compared with existing technologies, the ride-sharing matching method provided in this embodiment first performs hierarchical clustering on multiple groups of origin-destination (OD) locations to achieve preliminary clustering. Then, it further optimizes the clustering results by using the directed line segments formed by the origin and destination of each OD group, the destination location, and the travel time between the destinations of every two OD groups to improve the ride-sharing matching effect and thus enhance the user's ride-sharing experience. This invention performs ride-sharing matching based on actual OD locations, ensuring the effectiveness of ride-sharing matching and improving the user's ride-sharing experience.

[0056] For example, the step of performing hierarchical clustering on the multiple groups of ODs based on their starting positions to obtain a first clustering result includes:

[0057] Based on the starting positions of the multiple groups of ODs and the Birch algorithm, hierarchical clustering is performed on the multiple groups of ODs to obtain the first clustering result. The threshold parameter and classification factor of the Birch algorithm are determined based on the size of the OD data in different time periods.

[0058] Specifically, when performing hierarchical clustering of multiple ODs based on their starting positions, the Birch algorithm can be used to obtain the first clustering result. When using the Birch algorithm, the threshold parameter and classification factor can be determined based on the size of the OD data over different time periods. For example, experiments can be conducted beforehand using historical OD data to determine the threshold parameter and classification factor corresponding to different OD sizes, thus ensuring the effectiveness of the initial clustering. Throughout the clustering process, this invention does not specify the number of clusters, thereby achieving adaptive adjustment of the number of clusters to better adapt to the inherent structural characteristics of the dataset.

[0059] For example, the optimization of the first clustering result based on the directed line segments formed by the start and end points of each OD group within each cluster and the position of the end point in the first clustering result to obtain the second clustering result includes:

[0060] Based on the first clustering result, the angle between the directed line segments formed by the start and end points of every two groups of OD within each cluster, and the distance between the end points, the first clustering result is optimized to obtain the second clustering result.

[0061] Specifically, when optimizing the first clustering result based on the directed line segments formed by the start and end points of each OD group within each cluster and the position of the end points, the second clustering result can be obtained by considering the angle between the directed line segments formed by the start and end points of every two OD groups within each cluster and the distance between the end points. By using the angle between the directed line segments formed by the start and end points of every two OD groups within each cluster and the distance between the end points, the proximity between the paths corresponding to the two OD groups can be determined, thereby further optimizing the initial clustering result.

[0062] For example, the optimization of the first clustering result based on the angle between the directed line segments formed by the start and end points of every two groups of ODs within each cluster, and the distance between the end point positions, to obtain the second clustering result includes:

[0063] Based on the first clustering result, the angle between the directed line segments formed by the start and end points of every two groups of OD within each cluster, and the distance between the end points, the degree of opening and closing between every two groups of OD within each cluster is determined.

[0064] In the first clustering result, ODs within each cluster whose openness to any other OD is greater than a preset threshold are removed to obtain the second clustering result.

[0065] Specifically, when optimizing the first clustering result based on the angle between the directed line segments formed by the start and end points of every two OD groups within each cluster, and the distance between the end points, the opening degree between every two OD groups within each cluster can be determined. Then, ODs within each cluster whose opening degree with any other OD is greater than a preset threshold are removed, resulting in the second clustering result. By setting the opening degree, the proximity between the paths corresponding to two OD groups can be accurately reflected, thereby improving the multiplication matching effect.

[0066] For example, the degree of opening between every two groups of ODs within each cluster is calculated based on the following formula:

[0067]

[0068]

[0069] in, Indicates the degree of opening / closing between two sets of OD. and For weight parameters, This represents the angle between the directed line segments formed by the start and end points of two sets of origin (OD). This represents the distance between the endpoints of two sets of ODs. For the preset distance, The angle between the directed line segment formed by the start and end points of one of the two sets of ODs and the due north direction. It represents the angle between the direction of due north and the direction formed by the starting and ending points of the other OD in two OD groups.

[0070] Specifically, the degree of opening between any two groups of OD within each cluster can be determined using the formula described above.

[0071] For example, the optimization of the second clustering result based on the travel time between the destinations of every two groups of ODs within each cluster to obtain the third clustering result includes:

[0072] Based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, a similarity matrix is ​​constructed for each cluster.

[0073] Based on the similarity matrix corresponding to each cluster and the spectral clustering algorithm, the second clustering result is optimized to obtain the third clustering result.

[0074] Specifically, when optimizing the second clustering result based on the travel time between the destinations of every two OD groups within each cluster, a similarity matrix can be constructed for each cluster. This similarity matrix is ​​then used as the similarity matrix in the spectral clustering algorithm to optimize the second clustering result, resulting in the third clustering result. Constructing a similarity matrix for each cluster using the travel time between the destinations of every two OD groups within each cluster more accurately reflects the similarity between the paths corresponding to the two OD groups, thereby improving the performance of multi-path matching.

[0075] For example, the elements in the similarity matrix corresponding to each cluster are determined based on the following formula:

[0076]

[0077] in, This represents the similarity between any two groups of ODs within each cluster. This represents the travel time between the destinations of every two OD groups within each cluster. , , , , These are preset parameters. The first time threshold, This is the second time threshold.

[0078] Specifically, the elements in the similarity matrix corresponding to each cluster can be determined according to the above formula. and The value can be adjusted according to the actual situation; for example, it can be... Set it to 600 (in seconds). Set it to 1200. Also... , The settings need to ensure This ensures a smooth transition between intervals. The preset parameters can be adjusted based on the clustering accuracy obtained from the spectral clustering operation in the experiment, and the parameter with the highest accuracy should be selected as the preset parameter for practical application.

[0079] The technical solution of the present invention will be better illustrated below with a specific embodiment:

[0080] This invention utilizes the Birch algorithm and adjusts the threshold and branching factor parameters to perform adaptive hierarchical clustering on ride-hailing origin-destination (OD) data to initially divide OD point clusters. Directed line segments are constructed, and the angle and distance between OD vectors are calculated for data cleaning to assess the similarity between data points and further refine the initial clustering results. Combining crawled geographic road data and information, a custom function is used to calculate the travel time between each OD point, thereby constructing a similarity matrix. A spectral clustering algorithm is then applied to the similarity matrix for cluster analysis to obtain the final multi-trip matching result.

[0081] Combination Figure 2 Let's take a look. Figure 2 This is a flowchart illustrating an embodiment of the preliminary clustering process provided by the present invention. The parameter settings of the method proposed in this invention are based on in-depth adjustment, testing, and summarization of the acquired actual data, ensuring the scientific validity and effectiveness of the parameter settings. When the data size is less than the limit of small-scale data determined through experiments, the Birch model threshold parameter is set to 'a', and the branching factor is set to 'i'. If the data size is within a defined medium-scale range, the threshold is adjusted to 'b', and the branching factor is set to 'j'. When the data size exceeds the defined medium-scale range, the threshold is further refined to 'c', and the branching factor is set to 'k' (to ensure a reasonable and aesthetically pleasing cluster tree, the values ​​of branching factors i, j, and k do not exceed 10; based on actual testing of datasets of different sizes, a value of 7-9 for the branching factor parameter is preferable). Adaptive adjustment of the number of clusters: Throughout the clustering process, the present invention does not specify the number of clusters, achieving adaptive adjustment of the number of clusters, thereby better adapting to the inherent structural characteristics of the dataset.

[0082] Combination Figure 3 Let's take a look. Figure 3 This is a flowchart illustrating an embodiment of the vector angle calculation process provided by the present invention. The angle calculation assigns an angle value to each OD group by calculating the angle between the directed line segments formed by each OD group and the due north direction, and also calculates the angle between these line segments. The function is applied sequentially to calculate the angle between each line segment and the due north direction, as well as its actual length, and these calculation results are stored in a new data column. During the angle calculation process, if the absolute value α of the angle difference between two directed line segments is less than 180°, then α is the angle between them; if the absolute value α of the angle difference between two directed line segments is greater than 180°, then the angle value obtained by 360°-α is used as the angle between them. A smaller angle indicates a higher degree of directional similarity.

[0083] Even with a small angle between two directed line segments, points D will still appear far apart on a given distance scale. Therefore, after calculating the angle, an opening / closing constraint is added to further refine the clustering results. Opening / closing is an index that comprehensively considers the differences between two line segments in both direction and distance. When calculating opening / closing, the calculated angle difference is first used to assess the similarity in direction between the two line segments, reflecting their consistency in orientation. Next, the distance between the starting or ending points of the two line segments is calculated to assess their differences in spatial location, integrating these two dimensions into a unified opening / closing index. The weight for directional difference is set as ω1, and the weight for distance difference is set as ω2, satisfying the condition ω1 + ω2 = 1. Based on this, the formula for calculating the opening degree C is ω1*(angle difference / 180) + ω2*(distance / maximum distance), where the angle difference is normalized by dividing by 180 to ensure its value falls within the range of [0, 1]; the distance is normalized by dividing by a preset maximum distance value to ensure that the opening degree value also falls within the range of [0, 1]. The weights can be adjusted according to actual needs, thus allowing for a more flexible assessment of the overall difference between two line segments.

[0084] Combination Figure 4 Let's take a look. Figure 4 This is a flowchart illustrating an embodiment of the opening / closing constraint calculation process provided by the present invention. The OSM package can be used to calculate the travel time between each point D, thus taking into account the actual geographical conditions of the roads, such as road congestion and speed limits, to ensure the travel time is as accurate as possible. The travel time for each road segment is calculated as follows: Travel time for a road segment = (Congestion delay coefficient of that road segment × Road length) / Road speed limit. The road congestion delay coefficient is assigned by crawling hourly congestion data for different days of the week from a map open platform. The road speed limit can be obtained from a traffic platform. The total travel time is the sum of the travel times for each road segment. After obtaining the travel time between each point D, a travel time matrix can be constructed.

[0085] Combination Figure 5 Let's take a look. Figure 5 This is a flowchart illustrating an embodiment of the spectral clustering process provided by the present invention. Since the input data, i.e., the travel time, has a large dynamic range ([0, +∞), and the goal is to map it to a similarity scale [0, 1] with a negative correlation, and also to ensure that the portion of the travel time between 60s and 600s occupies a large portion of the data range in [0, 1], thereby improving clustering accuracy; simultaneously, the mapping must also ensure that when x=0, y=1. Through multiple experiments, the empirical formula is obtained as follows:

[0086]

[0087] The empirical formula is specifically designed to accurately map travel time (ranging from 0 to positive infinity) to similarity (ranging from 0 to 1). It deeply considers the broad dynamic range of travel time and strives to allocate a larger proportion of data within specific travel time ranges in the similarity space, thereby improving the accuracy of cluster analysis. The formula is presented as a piecewise function, meticulously dividing three intervals: from 0 to... Within a 600-second interval, using The logarithmic function form is used to slow down the rate at which similarity decreases with increasing driving time, thereby achieving higher resolution in shorter driving times (especially between 60 and 600 seconds); to (For example, twice as much) Within the interval (i.e., 1200 seconds), the function transforms into The linear form of the function is used to continue reducing similarity over longer travel times, but at a more gradual rate than the logarithmic part. The determination of d and e guarantees the function's stability over longer travel times. Continuity at the point and The conditions ensure a smooth transition between sections; when the travel time exceeds When the time interval is 0, the function value is always 0, meaning that the similarity between the two paths is extremely low and they can be considered dissimilar. The parameter selection aims to adjust the shape and position of the function to obtain an ideal similarity value within the target travel time range, and to ensure that the similarity is 1 when the travel time is 0. The goal of this mapping function is to occupy a large proportion of the similarity space within the travel time range of 60 seconds to 600 seconds, so as to finely distinguish the differences within this time range, while realizing a negative correlation mapping between travel time and similarity, and effectively transforming travel time into a limited similarity range, thereby optimizing the overall effect of cluster analysis.

[0088] This invention integrates hierarchical clustering and spectral clustering, and significantly improves the accuracy and efficiency of carpooling matching based on detailed analysis of OD points. Furthermore, by comprehensively considering various factors such as travel time and road congestion, it provides users with a more intelligent and personalized carpooling solution.

[0089] This invention also provides a ride-hailing matching device, combined with... Figure 6 Let's take a look. Figure 6 This is a schematic diagram of an embodiment of the ride-hailing matching device provided by the present invention. The ride-hailing matching device 600 includes:

[0090] Clustering module 601 is used to perform hierarchical clustering on the multiple groups of ODs based on the starting positions of the multiple groups of ODs to obtain a first clustering result;

[0091] The first optimization module 602 is used to optimize the first clustering result based on the directed line segment formed by the start and end points of each OD group in each cluster and the position of the end point in the first clustering result, so as to obtain the second clustering result.

[0092] The second optimization module 603 is used to optimize the second clustering result based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, to obtain a third clustering result, and to use the third clustering result as the ride-sharing matching result.

[0093] The specific implementation methods of each module of the ride-hailing matching device can be found in the description of the above-mentioned ride-hailing matching method, which has similar beneficial effects and will not be repeated here.

[0094] This invention also provides an electronic device, combined with Figure 7 Let's take a look. Figure 7 This is a schematic diagram of an embodiment of the electronic device provided by the present invention. The electronic device 700 includes a processor 701, a memory 702, and a computer program stored in the memory 702 and executable on the processor 701. When the processor 701 executes the program, it implements the ride-hailing matching method described above.

[0095] In a preferred embodiment, the electronic device 700 further includes a display 703 for displaying the processor 701 executing the ride-hailing matching method as described above.

[0096] For example, a computer program can be divided into one or more modules / units, one or more of which are stored in memory 702 and executed by processor 701 to complete the present invention. One or more modules / units can be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in electronic device 700. For example, the computer program can be divided into clustering module 601, first optimization module 602, and second optimization module 603 as described in the above embodiments. The specific functions of each module are as described above and will not be repeated here.

[0097] Electronic device 700 can be a desktop computer, laptop, PDA, or smartphone with an adjustable camera module.

[0098] The processor 701 may be an integrated circuit chip with signal processing capabilities. The processor 701 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0099] The memory 702 may be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory 702 stores programs, and the processor 701 executes these programs upon receiving execution instructions. The process definition method disclosed in any of the foregoing embodiments of the present invention can be applied to the processor 701, or implemented by the processor 701.

[0100] The display 703 can be an LCD screen or an LED screen. For example, the display screen on a mobile phone.

[0101] Understandable, Figure 7 The structure shown is only a schematic diagram of one possible structure of electronic device 700. Electronic device 700 may also include more than one of the following: Figure 7 Show more or fewer components. Figure 7 The components shown can be implemented using hardware, software, or a combination thereof.

[0102] The electronic device provided by the above embodiments of the present invention can be implemented with reference to the content specifically described in the above-described ride-hailing carpooling matching method, and has similar beneficial effects as the above-described ride-hailing carpooling matching method, which will not be repeated here.

[0103] This invention also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the ride-hailing matching method described above.

[0104] Generally, computer instructions for implementing the methods of the present invention can be carried on any combination of one or more computer-readable storage media. Non-transitory computer-readable storage media can include any computer-readable medium except for signals themselves that are temporarily propagating.

[0105] Computer-readable storage media can be, for example—but not limited to—electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or any combination thereof. More specific examples (a non-exhaustive list) of computer-readable storage media include: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in connection with an instruction execution system, apparatus, or device.

[0106] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. In particular, Python, suitable for neural network computation, and platform frameworks such as TensorFlow and PyTorch can be used. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0107] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware, and the program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0108] This invention discloses a ride-sharing matching method, device, electronic device, and storage medium. First, it performs hierarchical clustering on multiple groups of origin-destination (OD) locations to achieve preliminary clustering. Then, it further optimizes the clustering results by using the directed line segments formed by the origin and destination of each OD group, the destination location, and the travel time between the destinations of any two OD groups, thereby improving the ride-sharing matching effect and enhancing the user's ride-sharing experience. This invention performs ride-sharing matching based on actual OD locations, ensuring the effectiveness of the matching and improving the user's ride-sharing experience.

[0109] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A ride-hailing carpooling matching method, characterized in that, include: Based on the starting positions of multiple groups of ODs, hierarchical clustering is performed on the multiple groups of ODs to obtain the first clustering result; Based on the directed line segments formed by the start and end points of each OD group within each cluster in the first clustering result, and the position of the end point, the first clustering result is optimized to obtain the second clustering result. Based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, the second clustering result is optimized to obtain a third clustering result, and the third clustering result is used as the ride-sharing matching result; Based on the first clustering result, the angle between the directed line segments formed by the start and end points of every two groups of OD within each cluster, and the distance between the end points, the degree of opening and closing between every two groups of OD within each cluster is determined. In the first clustering result, ODs whose opening degree with any other OD within each cluster is greater than a preset threshold are removed to obtain the second clustering result; The degree of opening between every two groups of ODs within each cluster is calculated based on the following formula: in, Indicates the degree of opening / closing between two sets of OD. and For weight parameters, This represents the angle between the directed line segments formed by the start and end points of two sets of origin (OD). This represents the distance between the endpoints of two sets of ODs. For the preset distance, The angle between the directed line segment formed by the start and end points of one of the two sets of ODs and the due north direction. It represents the angle between the direction of due north and the direction formed by the starting and ending points of the other OD in two OD groups; Based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, a similarity matrix is ​​constructed for each cluster. Based on the similarity matrix corresponding to each cluster and the spectral clustering algorithm, the second clustering result is optimized to obtain the third clustering result; The elements in the similarity matrix corresponding to each cluster are determined based on the following formula: in, This represents the similarity between any two groups of ODs within each cluster. This represents the travel time between the destinations of every two OD groups within each cluster. , , , , These are preset parameters. The first time threshold, This is the second time threshold.

2. The ride-hailing matching method according to claim 1, characterized in that, The first clustering result is obtained by hierarchically clustering the multiple groups of ODs based on their starting positions, including: Based on the starting positions of the multiple groups of ODs and the Birch algorithm, hierarchical clustering is performed on the multiple groups of ODs to obtain the first clustering result. The threshold parameter and classification factor of the Birch algorithm are determined based on the size of the OD data in different time periods.

3. A ride-hailing matching device, characterized in that, include: The clustering module is used to perform hierarchical clustering of the multiple groups of ODs based on their starting positions to obtain the first clustering result; The first optimization module is used to optimize the first clustering result based on the directed line segment formed by the start and end points of each OD group in each cluster and the position of the end point in the first clustering result, so as to obtain the second clustering result. The second optimization module is used to optimize the second clustering result based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, to obtain a third clustering result, and to use the third clustering result as the ride-sharing matching result; Based on the first clustering result, the angle between the directed line segments formed by the start and end points of every two groups of OD within each cluster, and the distance between the end points, the degree of opening and closing between every two groups of OD within each cluster is determined. In the first clustering result, ODs whose opening degree with any other OD within each cluster is greater than a preset threshold are removed to obtain the second clustering result; The degree of opening between every two groups of ODs within each cluster is calculated based on the following formula: in, Indicates the degree of opening / closing between two sets of OD. and For weight parameters, This represents the angle between the directed line segments formed by the start and end points of two sets of origin (OD). This represents the distance between the endpoints of two sets of ODs. For the preset distance, The angle between the directed line segment formed by the start and end points of one of the two sets of ODs and the due north direction. It represents the angle between the direction of due north and the direction formed by the starting and ending points of the other OD in two OD groups; Based on the travel time between the destinations of every two groups of ODs within each cluster in the second clustering result, a similarity matrix is ​​constructed for each cluster. Based on the similarity matrix corresponding to each cluster and the spectral clustering algorithm, the second clustering result is optimized to obtain the third clustering result; The elements in the similarity matrix corresponding to each cluster are determined based on the following formula: in, This represents the similarity between any two groups of ODs within each cluster. This represents the travel time between the destinations of every two OD groups within each cluster. , , , , These are preset parameters. The first time threshold, This is the second time threshold.

4. An electronic device, characterized in that, It includes a memory and a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, it implements the ride-hailing matching method according to claim 1 or 2.

5. A computer-readable storage medium, characterized in that, It stores a computer program, which, when executed by a processor, implements the ride-hailing matching method as described in claim 1 or 2.

Images