A vehicle dispatching method, device, apparatus and storage medium

Abstract

The application relates to the technical field of online car hailing, in particular to a vehicle dispatching method, device and equipment and a storage medium, which are used for solving the problem of low dispatching success rate. The method comprises the following steps: based on vehicle feature information in each grid, determining a to-be-dispatched vehicle from vehicles currently located in a corresponding grid, and obtaining associated information of each to-be-dispatched vehicle; then, based on the associated information of each to-be-dispatched vehicle, predicting an acceptance probability of a driver of each to-be-dispatched vehicle to each to-be-dispatched task; and based on each acceptance probability and a preset screening rule, determining a target dispatching vehicle matched with each target grid from all to-be-dispatched vehicles, wherein the target grid comprises at least one dispatching demand; in this way, the target dispatching vehicle can be screened from all to-be-dispatched vehicles according to the preset screening rule, so that the to-be-dispatched task is preferentially sent to the target dispatching vehicle, thereby improving the dispatching success rate.

Description

Technical Field

[0001] This application relates to the field of ride-hailing technology, and in particular to a method, apparatus, device and storage medium for vehicle dispatching. Background Technology

[0002] In actual ride-hailing operations, supply and demand imbalances often occur in certain areas. In such cases, ride-hailing platform servers typically allocate resources by dispatching redundant vehicles from areas with lower demand to areas with higher demand.

[0003] In related technologies, ride-hailing platform servers often notify redundant vehicles of dispatch tasks by pushing these tasks to their in-vehicle terminals in areas with fewer orders. This dispatch information includes, but is not limited to, the destination and estimated mileage. Drivers in these redundant vehicles can then accept or reject the dispatch task via their in-vehicle terminals after receiving it from the ride-hailing platform server.

[0004] However, through statistical analysis of drivers' behavior, the inventors discovered that different drivers have varying levels of acceptance of dispatch tasks at different times and / or in different areas, resulting in a low overall dispatch success rate for ride-hailing platform servers. Summary of the Invention

[0005] This application provides a method, apparatus, device, and storage medium for vehicle scheduling, in order to improve the overall scheduling success rate of the platform server.

[0006] The specific technical solutions provided in this application are as follows:

[0007] In a first aspect, embodiments of this application provide a method for vehicle dispatching, including:

[0008] Based on the predicted information of each grid within a preset time period in the future, the vehicle feature information of each grid is obtained;

[0009] Based on the vehicle feature information within each grid, vehicles to be dispatched are determined from the vehicles currently located in the corresponding grid, and the associated information of each vehicle to be dispatched is obtained.

[0010] Based on the associated information of each vehicle to be dispatched, predict the probability that the driver of each vehicle to be dispatched will accept each dispatched task.

[0011] Based on the acceptance probabilities and preset screening rules, target vehicles that match each target grid are determined from all vehicles to be dispatched, wherein the target grid includes at least one dispatching requirement.

[0012] The vehicle dispatching method provided in this application, compared with the related technology of directly pushing dispatching tasks to each redundant vehicle, predicts the acceptance probability of each driver of each vehicle for each dispatching task based on the association information of each vehicle before distributing the dispatching task to the vehicles to be dispatched. This allows for the prediction of the acceptance probability of each vehicle for each dispatching task in advance. Then, based on the filtering rules and the acceptance probabilities, each target dispatching vehicle is selected from multiple vehicles to be dispatched. In this way, the dispatching task is sent to the target dispatching vehicle first, improving the dispatching success rate of dispatching task distribution.

[0013] In one possible implementation, determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid includes:

[0014] For any grid in each grid, based on the number of redundant vehicles included in the vehicle feature information in any grid, each target vehicle is determined from the vehicles currently located in any grid, wherein the target vehicle includes at least one of empty vehicles and passenger vehicles whose remaining journey time is less than a preset time.

[0015] If the total number of all target vehicles is not greater than the number of redundant vehicles, then each target vehicle is determined as a vehicle to be dispatched.

[0016] If the total number of all target vehicles is greater than the number of redundant vehicles, then the target vehicles are sorted, and the vehicles to be dispatched are determined from the target vehicles based on the sorting results.

[0017] In one possible implementation, after determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid, and before obtaining the association information of each vehicle to be dispatched, the method further includes:

[0018] Based on the predicted information of each grid within a preset time period in the future, the demand characteristic information within each target grid is obtained;

[0019] Based on the demand feature information within each target grid and each vehicle to be dispatched, a task to be dispatched corresponding to each dispatch demand is generated, wherein the destination of the task to be dispatched is any point of interest (POI) in the target grid where the corresponding dispatch demand is located.

[0020] In one possible implementation, the associated information includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, and real-time characteristics of the driver of the vehicle to be dispatched; and the grid characteristics of the grid where the vehicle to be dispatched is located.

[0021] The step of obtaining the associated information for each vehicle to be dispatched includes:

[0022] Obtain the professional characteristics of the driver of each vehicle to be dispatched; and / or, obtain the historical behavioral characteristics of the driver of each vehicle to be dispatched; and / or, obtain the real-time characteristics of the driver of each vehicle to be dispatched; and / or, obtain the grid characteristics of the grid where each vehicle to be dispatched is located.

[0023] By using the above method, after determining each vehicle to be dispatched, multi-dimensional information about each vehicle is obtained, and based on this information, the acceptance probability of each vehicle for each dispatched task is predicted. This makes the predicted acceptance probability more accurate and more in line with the driver's wishes, thus sending the dispatched task to the vehicle with the higher acceptance probability and improving the dispatch success rate.

[0024] In one possible implementation, the historical behavioral characteristics of the driver of any vehicle to be dispatched are obtained in the following way:

[0025] The first time interval is obtained based on the last dispatch task push time and the current time of any of the vehicles to be dispatched;

[0026] The second time interval is obtained based on the last time the dispatch task was rejected and the current time for any of the vehicles to be dispatched;

[0027] The task acceptance success rate of any vehicle to be dispatched, the average task acceptance time and the average task rejection time of the driver of any vehicle to be dispatched are statistically analyzed within the first preset historical time period.

[0028] The first time interval, the second time interval, the task acceptance success rate, the average task acceptance time, and the average task rejection time are used as part or all of the historical behavioral feature information of the driver of any vehicle to be dispatched.

[0029] By using the above method, we can obtain the historical behavioral characteristics of the drivers of each vehicle to be dispatched. This allows us to more accurately predict the probability of the drivers of each vehicle accepting the dispatched task corresponding to each dispatch request based on the historical behavioral characteristics.

[0030] In one possible implementation, the real-time characteristic information of the driver of any vehicle to be dispatched is obtained in the following way:

[0031] Based on the current time, determine the time identifier of the task to be scheduled corresponding to any scheduling request;

[0032] Based on the current location of any vehicle to be dispatched and the destination of the task to be dispatched corresponding to any dispatch request, the estimated arrival time is obtained.

[0033] Based on the target time and the current time of any vehicle to be dispatched, the empty driving time is obtained, wherein the target time includes at least one of the last completed order time and the online time of the day;

[0034] Based on the online time of any vehicle to be dispatched on the same day and the current time, the online duration on the same day is obtained, and the number of completed orders of any vehicle to be dispatched within the second preset historical duration is counted.

[0035] The time identifier, the empty driving time, the estimated arrival time, the online time of the day, and the number of completed orders are used as some or all of the real-time characteristic information of the driver of any vehicle to be dispatched.

[0036] By using the above method, real-time characteristic information of the driver of each vehicle to be dispatched can be obtained. In this way, by combining the real-time situation of the driver of each vehicle to be dispatched, the probability of the driver of the vehicle to be dispatched accepting the dispatched task corresponding to each dispatch request can be predicted in a targeted manner, so as to improve the accuracy of the predicted acceptance probability and thus improve the dispatch success rate.

[0037] In one possible implementation, the grid feature information of any grid containing a vehicle to be scheduled is obtained as follows:

[0038] The acceptance rate, rejection rate, and cancellation rate of the target scheduling task in the grid where any vehicle to be scheduled is located within the third preset historical time period are statistically analyzed. The starting grid of the target scheduling task is the grid where any vehicle to be scheduled is located, and the destination grid is the target grid where the corresponding scheduling request is located.

[0039] Obtain the reference grid information of the grid where the driver of any vehicle to be dispatched is located when the driver accepts the target dispatch task, and the destination grid information of the destination grid;

[0040] The acceptance rate, rejection rate, cancellation rate, reference grid information, and destination grid information are used as part or all of the grid feature information of the grid where any vehicle to be dispatched is located.

[0041] Using the above method, we can obtain the grid feature information of the grid where each vehicle to be dispatched is located. This allows for a more comprehensive prediction of the acceptance probability of the vehicle to be dispatched, and makes the subsequent prediction of the driver's acceptance probability of each dispatch request based on the grid feature information more accurate, thereby improving the dispatch success rate.

[0042] In one possible implementation, predicting the probability of each driver accepting each dispatched task based on the association information of each vehicle to be dispatched includes:

[0043] Input the associated information of each vehicle to be dispatched into the target prediction model to obtain the probability of each driver of each vehicle accepting each dispatched task.

[0044] The target prediction model is trained based on a prediction training set, which includes multiple training samples. Each training sample includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, real-time characteristics of any historical scheduling task, and grid characteristics of the vehicle corresponding to the training sample, as well as the driver's acceptance information for any historical scheduling task.

[0045] The above method, by pre-training the target prediction model, enables the rapid calculation of the acceptance probability of a vehicle to be dispatched for a dispatched task corresponding to a certain dispatching demand during the implementation process, thereby improving the overall dispatching success rate of the platform server.

[0046] In one possible implementation, the filtering rules include a scheduling objective function and task distribution constraints; the step of determining the target scheduling vehicle matching each target grid from all vehicles to be scheduled based on each acceptance probability and the preset filtering rules includes:

[0047] Based on the aforementioned acceptance probabilities, the success probability of scheduling the task to be scheduled corresponding to each scheduling request is determined.

[0048] Based on the scheduling success probability of the task to be scheduled corresponding to each scheduling request, a preset optimization algorithm is invoked to determine each target matching pair so that the scheduling objective function takes the optimal value under the task distribution constraint. The scheduling objective function represents the average scheduling success probability of the task to be scheduled corresponding to all scheduling requests. The task to be scheduled corresponding to the scheduling request in the target matching pair is distributed to the vehicle to be scheduled in the target matching pair.

[0049] The vehicles to be scheduled in each target matching pair are respectively determined as target scheduling vehicles that match the target grid where the corresponding scheduling requirement is located.

[0050] The above method, by setting a scheduling objective function and task distribution constraints, can determine the optimal scheduling task distribution combinations (i.e., target matching pairs) from multiple vehicles to be scheduled, which maximize the average scheduling success probability of the tasks corresponding to all scheduling requests. Based on these scheduling task distribution combinations, the tasks corresponding to all scheduling requests are distributed, thereby improving the scheduling success rate.

[0051] Secondly, embodiments of this application provide a vehicle dispatching apparatus, comprising:

[0052] The feature information prediction module is used to obtain vehicle feature information in each grid based on the prediction information of each grid within a preset time period in the future;

[0053] The acquisition module is used to determine the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information in each grid, and to acquire the association information of each vehicle to be dispatched.

[0054] The probability prediction module is used to predict the probability that the driver of each vehicle to be dispatched will accept each dispatched task based on the association information of each vehicle to be dispatched.

[0055] The determination module is used to determine the target dispatch vehicle that matches each target grid from all vehicles to be dispatched based on each acceptance probability and preset screening rules, wherein the target grid includes at least one dispatch requirement.

[0056] In one possible implementation, the acquisition module is used for:

[0057] For any grid in each grid, based on the number of redundant vehicles included in the vehicle feature information in any grid, each target vehicle is determined from the vehicles currently located in any grid, wherein the target vehicle includes at least one of empty vehicles and passenger vehicles whose remaining journey time is less than a preset time.

[0058] If the total number of all target vehicles is not greater than the number of redundant vehicles, then each target vehicle is determined as a vehicle to be dispatched.

[0059] If the total number of all target vehicles is greater than the number of redundant vehicles, then the target vehicles are sorted, and the vehicles to be dispatched are determined from the target vehicles based on the sorting results.

[0060] In one possible implementation, after determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid, and before obtaining the association information of each vehicle to be dispatched, the acquisition module is further configured to:

[0061] Based on the predicted information of each grid within a preset time period in the future, the demand characteristic information within each target grid is obtained;

[0062] Based on the demand feature information within each target grid and each vehicle to be dispatched, a task to be dispatched corresponding to each dispatch demand is generated, wherein the destination of the task to be dispatched is any point of interest (POI) in the target grid where the corresponding dispatch demand is located.

[0063] In one possible implementation, the associated information includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, and real-time characteristics of the driver of the vehicle to be dispatched; and the grid characteristics of the grid where the vehicle to be dispatched is located.

[0064] The acquisition module is specifically used to: acquire the professional characteristics information of the driver of each vehicle to be dispatched; and / or acquire the historical behavioral characteristics information of the driver of each vehicle to be dispatched; and / or acquire the real-time characteristics information of the driver of each vehicle to be dispatched; and / or acquire the grid characteristics information of the grid where each vehicle to be dispatched is located.

[0065] In one possible implementation, the historical behavioral characteristics of the driver of any vehicle to be dispatched are obtained in the following way:

[0066] The first time interval is obtained based on the last dispatch task push time and the current time of any of the vehicles to be dispatched;

[0067] The second time interval is obtained based on the last time the dispatch task was rejected and the current time for any of the vehicles to be dispatched;

[0068] The task acceptance success rate of any vehicle to be dispatched, the average task acceptance time and the average task rejection time of the driver of any vehicle to be dispatched are statistically analyzed within the first preset historical time period.

[0069] The first time interval, the second time interval, the task acceptance success rate, the average task acceptance time, and the average task rejection time are used as part or all of the historical behavioral feature information of the driver of any vehicle to be dispatched.

[0070] In one possible implementation, the real-time characteristic information of the driver of any vehicle to be dispatched is obtained in the following way:

[0071] Based on the current time, determine the time identifier of the task to be scheduled corresponding to any scheduling request;

[0072] Based on the current location of any vehicle to be dispatched and the destination of the task to be dispatched corresponding to any dispatch request, the estimated arrival time is obtained.

[0073] Based on the target time and the current time of any vehicle to be dispatched, the empty driving time is obtained, wherein the target time includes at least one of the last completed order time and the online time of the day;

[0074] Based on the online time of any vehicle to be dispatched on the same day and the current time, the online duration on the same day is obtained, and the number of completed orders of any vehicle to be dispatched within the second preset historical duration is counted.

[0075] The time identifier, the empty driving time, the estimated arrival time, the online time of the day, and the number of completed orders are used as some or all of the real-time characteristic information of the driver of any vehicle to be dispatched.

[0076] In one possible implementation, the grid feature information of any grid containing a vehicle to be scheduled is obtained as follows:

[0077] The acceptance rate, rejection rate, and cancellation rate of the target scheduling task in the grid where any vehicle to be scheduled is located within the third preset historical time period are statistically analyzed. The starting grid of the target scheduling task is the grid where any vehicle to be scheduled is located, and the destination grid is the target grid where the corresponding scheduling request is located.

[0078] Obtain the reference grid information of the grid where the driver of any vehicle to be dispatched is located when the driver accepts the target dispatch task, and the destination grid information of the destination grid;

[0079] The acceptance rate, rejection rate, cancellation rate, reference grid information, and destination grid information are used as part or all of the grid feature information of the grid where any vehicle to be dispatched is located.

[0080] In one possible implementation, the probability prediction module is used for:

[0081] The associated information of each vehicle to be dispatched is input into the target prediction model to obtain the probability of each driver of each vehicle accepting each dispatched task.

[0082] The target prediction model is trained based on a prediction training set, which includes multiple training samples. Each training sample includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, real-time characteristics of any historical scheduling task, and grid characteristics of the vehicle corresponding to the training sample, as well as the driver's acceptance information for any historical scheduling task.

[0083] In one possible implementation, the filtering rules include a scheduling objective function and task distribution constraints; the determining module is used for:

[0084] Based on the aforementioned acceptance probabilities, the success probability of scheduling the task to be scheduled corresponding to each scheduling request is determined.

[0085] Based on the scheduling success probability of the task to be scheduled corresponding to each scheduling request, a preset optimization algorithm is invoked to determine each target matching pair so that the scheduling objective function takes the optimal value under the task distribution constraint. The scheduling objective function represents the average scheduling success probability of the task to be scheduled corresponding to all scheduling requests. The task to be scheduled corresponding to the scheduling request in the target matching pair is distributed to the vehicle to be scheduled in the target matching pair.

[0086] The vehicles to be scheduled in each target matching pair are respectively determined as target scheduling vehicles that match the target grid where the corresponding scheduling requirement is located.

[0087] Thirdly, embodiments of this application provide an electronic device, including:

[0088] Memory is used to store computer programs or instructions;

[0089] A processor for executing a computer program or instructions in the memory, such that the method described in any of the first aspects is performed.

[0090] Fourthly, embodiments of this application provide a computer-readable storage medium that, when instructions in the storage medium are executed by a processor, enables the processor to perform the method described in any one of the first aspects above.

[0091] Fifthly, embodiments of this application provide a computer program product comprising: computer program code, which, when executed on a computer, causes the computer to perform the method described in any one of the first aspects.

[0092] Furthermore, the technical effects of any of the implementation methods in the second to fifth aspects can be found in the technical effects of different implementation methods in the first aspect, and will not be repeated here.

[0093] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0094] Figure 1 This is a schematic diagram of an application scenario in an embodiment of this application;

[0095] Figure 2 This is a flowchart illustrating a vehicle dispatching method according to an embodiment of this application;

[0096] Figure 3 This is a flowchart illustrating a method for determining a vehicle to be dispatched according to an embodiment of this application;

[0097] Figure 4This is a flowchart illustrating a method for obtaining historical behavioral characteristic information of drivers of vehicles to be dispatched, as described in an embodiment of this application.

[0098] Figure 5 This is a flowchart illustrating a method for obtaining real-time characteristic information of drivers of vehicles to be dispatched, as described in an embodiment of this application.

[0099] Figure 6 This is a flowchart illustrating a method for obtaining grid feature information of the grid where a vehicle to be dispatched is located, according to an embodiment of this application.

[0100] Figure 7 This is a flowchart illustrating a method for determining a target dispatch vehicle according to an embodiment of this application;

[0101] Figure 8 This is a schematic diagram of the logical architecture of a vehicle dispatching device according to an embodiment of this application;

[0102] Figure 9 This is a schematic diagram of the physical architecture of the electronic device in the embodiments of this application. Detailed Implementation

[0103] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0104] It should be noted that the terms "first," "second," "third," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in sequences other than those illustrated or described herein.

[0105] In this embodiment of the application, in order to solve the problem of low scheduling success rate of ride-hailing platform servers under related technologies, this embodiment of the application provides a vehicle scheduling method. Specifically, based on the prediction information of each grid within a preset time period in the future, the vehicle feature information of each grid is obtained. Then, based on the vehicle feature information of each grid, the vehicles to be scheduled are determined from the vehicles currently located in the corresponding grid, and the association information of each vehicle to be scheduled is obtained. Then, based on the association information of each vehicle to be scheduled, the acceptance probability of each driver of each vehicle to be scheduled for each task is predicted. Finally, based on the acceptance probabilities and preset filtering rules, the target vehicles matching each target grid are determined from all vehicles to be scheduled. The target grid includes at least one scheduling request. In this way, the target vehicles to be scheduled can be selected from all vehicles to be scheduled according to the preset filtering rules, thereby prioritizing the sending of tasks to the target vehicles and improving the scheduling success rate of task issuance.

[0106] The preferred embodiments of this application will be further described in detail below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only for illustration and explanation of this application and are not intended to limit this application. Furthermore, the embodiments of this application and the features in the embodiments can be combined with each other without conflict.

[0107] See Figure 1 The diagram shown illustrates an application scenario of an embodiment of this application. The application scenario includes a user terminal 100, a server 200, and an in-vehicle terminal 300. The user terminal 100 is connected to the server 200 via a wired or wireless network, and the in-vehicle terminal 300 is also connected to the server 200 via a wired or wireless network.

[0108] User terminal 100 sends an order to server 200. The order can be initiated by the user through the order placement terminal of the ride-hailing platform server or by the user through the order placement terminal of a third-party platform server. The order includes order information such as origin location information, destination location information, user's current location information, and communication method.

[0109] The vehicle terminal 300 sends the vehicle's location information and order status to the server 200.

[0110] After receiving an order, server 200 determines the geographical area where the user is located based on the order information; and based on the received vehicle location information and order acceptance status, it identifies vehicles that have not yet accepted orders or are about to complete orders within the geographical area where the order information is located, thereby completing the order dispatch.

[0111] It should be noted that Figure 1The location of the vehicle-mounted terminal 300 in the diagram is for illustrative purposes only and does not represent a fixed location for actual installation in a vehicle. The vehicle-mounted terminal 300 can also be located in other parts of the vehicle, such as in the door or between the front seats.

[0112] After introducing the application scenarios of the embodiments of this application, the embodiments of this application provide a method for vehicle dispatching. It should be noted that this method is not limited to ride-hailing platform servers, but can also be applied to any system with similar resource dispatching needs. The technical solution of this application will be described in detail below using a ride-hailing platform server as an example. See also Figure 2 As shown, the specific process of the above method is as follows:

[0113] Step 400: Based on the prediction information of each grid within a preset time period in the future, obtain the vehicle feature information of each grid.

[0114] Here, a grid is any grid in a grid world that represents a real-world geographic region.

[0115] In this embodiment of the application, before executing step 400, the existing supply and demand forecasting module is first invoked to predict the supply and demand relationship of each grid in the area served by the ride-hailing platform server within a preset time period in the future, thereby obtaining the forecast information of each grid. Then, step 400 is executed to obtain the vehicle feature information in each grid based on the forecast information of each grid.

[0116] The vehicle feature information includes the number of redundant vehicles within the corresponding grid.

[0117] For example, suppose grid A is any grid within the service area of ​​the ride-hailing platform server, and the prediction information for grid A obtained through the supply and demand forecasting model includes: the number of demand orders in the next 15 minutes is 3, and the number of vehicles is 10. Then, based on the prediction information of grid A, the vehicle characteristic information of grid A in the next 15 minutes is obtained, and the number of redundant vehicles included in this vehicle characteristic information is (10-3) = 7.

[0118] In another example, suppose grid B is any grid within the service area of ​​the ride-hailing platform server, and the predicted information for grid B obtained through the supply and demand forecasting model includes: the number of demand orders in the next 15 minutes is 6, and the number of vehicles is 6. Then, based on the predicted information for grid B, the vehicle characteristic information for grid B in the next 15 minutes is obtained, and the number of redundant vehicles included in this vehicle characteristic information is (6-6) = 0. In this way, the existing supply and demand forecasting module within the ride-hailing platform server can predict the supply and demand relationship of each grid within the service area of ​​the ride-hailing platform server in the future within a preset time period. This allows for advance knowledge of supply and demand imbalances in certain grids within the future preset time period, and allows for the pre-allocation of vehicles from areas with fewer demand orders to areas with higher demand orders to balance resources among the grids. It should be noted that the above 15 minutes is merely an example, and this application does not specifically limit the above content.

[0119] Step 410: Based on the vehicle feature information in each grid, determine the vehicles to be dispatched from the vehicles currently located in the corresponding grid, and obtain the association information of each vehicle to be dispatched.

[0120] In the embodiments of this application, see the following: Figure 3 As shown, when determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid in step 410, the vehicles to be dispatched can be determined by executing the following steps:

[0121] Step 4101: For any of the above grids, based on the number of redundant vehicles included in the vehicle feature information of any grid, determine each target vehicle from the vehicles currently located in any grid, wherein the target vehicle includes at least one of empty vehicles and passenger vehicles whose remaining journey length is less than a preset value.

[0122] Step 4102: If the total number of the above target vehicles is not greater than the number of redundant vehicles, then each of the above target vehicles is determined as a vehicle to be dispatched.

[0123] Step 4103: If the total number of the above target vehicles is greater than the number of redundant vehicles, sort the above target vehicles and determine the vehicles to be dispatched from the target vehicles based on the sorting results.

[0124] In this embodiment of the application, when performing step 4103, if the total number of the above target vehicles is greater than the number of redundant vehicles, the above target vehicles can be sorted in descending order of empty running time. Then, based on the sorting result, the target vehicles with the number of redundant vehicles are selected from the target vehicles, and the selected target vehicles are determined as vehicles to be dispatched.

[0125] In this way, the above method can identify the vehicles to be dispatched in each grid. Then, based on the prediction information of each grid, the demand feature information of each target grid is obtained. Based on the demand feature information of each target grid and each vehicle to be dispatched, a dispatch task corresponding to each dispatch demand is generated. The destination of the dispatch task is any point of interest (POI) in the target grid where the corresponding dispatch demand is located.

[0126] The demand characteristic information includes the number of scheduling demands within the corresponding target grid.

[0127] For example, suppose grid C is any grid within the service area of ​​the ride-hailing platform server, and the prediction information for grid C obtained through the supply and demand forecasting model includes: the number of demand orders in the next 15 minutes is 10, and the number of vehicles is 5. Then, based on the prediction information of grid C, the demand characteristic information of grid C (i.e., the target grid) in the next 15 minutes is obtained, and the number of scheduling demands included in this demand characteristic information is (10-5) = 5.

[0128] In practical applications, POIs in the grid are usually used as the destination of scheduling tasks. When there are multiple POIs in a grid, the vehicle to be scheduled is usually scheduled to the POI in the grid that is closest to the vehicle.

[0129] After obtaining the task to be scheduled corresponding to each scheduling request, step 410 is executed to obtain the associated information of each vehicle to be scheduled. The associated information includes at least one of the following: the professional characteristics information, historical behavior information, and real-time characteristics information of the driver of the vehicle to be scheduled, and the grid information of the grid where the vehicle to be scheduled is located.

[0130] In this embodiment of the application, when obtaining the associated information of each vehicle to be dispatched in step 410, some or all of the following operations one to four can be performed:

[0131] Operation 1: Obtain the occupational characteristics information of the drivers of each vehicle to be dispatched.

[0132] In practice, the aforementioned occupational characteristics information can be obtained by acquiring the basic information of the driver of the corresponding vehicle to be dispatched. The occupational characteristics information includes some or all of the driver's age, gender, and number of days employed.

[0133] Step 2: Obtain the historical behavioral characteristics of the drivers of each vehicle to be dispatched.

[0134] For specific implementation details, please refer to... Figure 4As shown, the historical behavioral characteristics of the driver of any vehicle to be dispatched can be obtained by performing the following steps:

[0135] Step 500: Based on the last dispatch task push time and the current time of any of the above-mentioned vehicles to be dispatched, obtain the first time interval.

[0136] Step 510: Based on the previous dispatch task rejection time and the current time of any of the above-mentioned vehicles to be dispatched, obtain the second time interval.

[0137] Step 520: Calculate the task acceptance success rate of any of the above-mentioned vehicles to be dispatched, the average task acceptance time of any of the above-mentioned vehicles to be dispatched, and the average task rejection time of the drivers of any of the above-mentioned vehicles to be dispatched within the first preset historical time period.

[0138] The task reception success rate within the first preset historical period can be the task reception success rate over the past k days, where k = 1, 3, 7, 15, 30...

[0139] Step 530: Use some or all of the following as historical behavioral characteristic information of the driver of any of the above-mentioned vehicles to be dispatched: the first time interval, the second time interval, the task acceptance success rate, the average task acceptance time, and the average task rejection time.

[0140] In this way, by performing steps 500 to 530 above, the historical behavioral characteristics information of the drivers of each vehicle to be dispatched can be obtained.

[0141] Step 3: Obtain real-time characteristic information of the driver of each vehicle to be dispatched.

[0142] For specific implementation details, please refer to... Figure 5 As shown, the real-time characteristic information of the driver of any vehicle to be dispatched is obtained by performing the following operations:

[0143] Step 500': Based on the current time, determine the time identifier of the task to be scheduled corresponding to any of the above scheduling requests.

[0144] The time identifier represents the time period in which the task to be scheduled corresponds to any of the above scheduling requests.

[0145] Step 510': Based on the current location of any of the vehicles to be dispatched and the destination of the task to be dispatched corresponding to any of the dispatch requests, obtain the estimated arrival time.

[0146] Step 520': Based on the target time and current time of any of the above-mentioned vehicles to be dispatched, obtain the empty driving time, wherein the target time includes at least one of the last order completion time and the online time of the day.

[0147] In real-world scenarios, considering the possibility that a driver might not receive any orders after going online, when executing step 520', if any driver of any of the aforementioned vehicles awaiting dispatch has not received any orders since going online that day, the empty driving time is calculated based on the online time of that day and the current time. The online time of that day can refer to the most recent online time of any of the aforementioned vehicles awaiting dispatch that day. The last order completion time can refer to the most recent order completion time (from the current time) completed by any of the aforementioned vehicles awaiting dispatch that day.

[0148] In some embodiments, if the target time for vehicle A to be dispatched includes the last order completion time and the online time of the current day, then the time closest to the present (current time) is selected from the last order completion time and the online time of the current day, and that time is used as the target time for vehicle A to be dispatched.

[0149] Step 530': Based on the online time of any of the above-mentioned vehicles to be dispatched on the same day and the current time, obtain the online duration on the same day, and count the number of completed orders of any of the above-mentioned vehicles to be dispatched within the second preset historical duration.

[0150] The number of completed orders within the second preset historical time period can be the number of completed orders in the past hour.

[0151] Step 540': Use some or all of the following as the real-time characteristic information of the driver of any of the above-mentioned vehicles to be dispatched: time stamp, empty driving time, estimated arrival time, online time on the day, and number of completed orders.

[0152] In this way, by performing the above steps 500' to 540', the real-time characteristic information of the driver of each vehicle to be dispatched can be obtained.

[0153] Step 4: Obtain the grid feature information of the grid where each vehicle to be dispatched is located.

[0154] For specific implementation details, please refer to... Figure 6 As shown, the grid feature information of any grid containing a vehicle to be dispatched is obtained by performing the following operations:

[0155] Step 500: Calculate the acceptance rate, rejection rate, and cancellation rate of the target scheduling task in the grid where any of the above-mentioned vehicles to be scheduled are located within the third preset historical time period. The starting grid of the target scheduling task is the grid where any of the above-mentioned vehicles to be scheduled are located, and the destination grid is the target grid where the corresponding scheduling request is located.

[0156] Among them, the task reception success rate within the third preset historical time period can be the task reception success rate of the past K days, where K = 1, 3, 7, 15, 30.

[0157] Step 510: Obtain the reference grid information of the grid where the driver of any of the above-mentioned vehicles to be dispatched is located and the destination grid information of the destination grid when the driver accepts the target dispatch task.

[0158] Step 520: Take some or all of the acceptance rate, rejection rate, cancellation rate, reference grid information and destination grid information as the grid feature information of any of the above-mentioned vehicles to be dispatched.

[0159] Thus, by executing steps 500” to 520” above, the grid feature information of each grid where the vehicle to be dispatched is located can be obtained.

[0160] Step 420: Based on the association information of each vehicle to be dispatched, predict the probability that the driver of each vehicle to be dispatched will accept each dispatched task.

[0161] In this embodiment of the application, before implementing the above-described vehicle scheduling method, a prediction training set is collected in advance. The prediction training set includes multiple training samples. The training samples include at least one of the following: the driver's professional characteristics information, historical behavior characteristics information, real-time characteristics information of any historical scheduling task, and grid characteristics information of the driver of the vehicle corresponding to the training sample, as well as the driver's acceptance information for any historical scheduling task.

[0162] It should be noted that the occupational characteristics, historical behavior characteristics, real-time characteristics, and grid characteristics of each training sample can be obtained in the same way as the occupational characteristics, historical behavior characteristics, real-time characteristics, and grid characteristics of any of the aforementioned vehicles to be dispatched, and will not be elaborated here.

[0163] Then, after collecting the above-mentioned prediction training set, the prediction model to be trained is subjected to multiple rounds of iterative training based on the above-mentioned prediction training set until the prediction model to be trained meets the convergence condition. The prediction model after the last round of training is used as the above-mentioned target prediction model.

[0164] In this embodiment of the application, no limitations are placed on the construction method or model structure of the target prediction model. For example, the target prediction model can be a model built based on any machine learning algorithm, such as XGBoost or LightGBM; or it can be a model built based on any deep learning algorithm, such as DeepFM.

[0165] Furthermore, this application does not impose specific limitations on the model training process and the selection method of the loss function, and any model training method can be used to achieve the above-mentioned model training function.

[0166] In some embodiments, the target prediction model can also be optimized and trained, that is, subsequent data is collected, the subsequent data is processed to obtain a new prediction sample set, and then the target prediction model is optimized using the new prediction sample set. The specific optimization method can be a general model optimization method, which will not be elaborated here.

[0167] After obtaining the trained target prediction model, in the process of implementing the vehicle scheduling method provided in this application, when executing step 420, the associated information of each vehicle to be scheduled can be specifically input into the target prediction model to obtain the acceptance probability of each driver of each vehicle to be scheduled for each task to be scheduled.

[0168] For example, suppose there are M vehicles to be dispatched and N dispatch requests. Then, in step 420, the association information x of the vehicle to be dispatched j (1≤j≤M) with the dispatch request i (1≤i≤N) is... ij Input the target prediction model, and the output of the target prediction model is the probability that the vehicle j to be scheduled will accept the task to be scheduled corresponding to the scheduling demand i, such as 80%.

[0169] In this embodiment of the application, it is still assumed that there are M vehicles to be dispatched and N dispatch requests. Then, the probability that the driver of vehicle j will accept the dispatch request corresponding to dispatch request i can be expressed by the following formula:

[0170] p ij =p(y=1|x ij )

[0171] Where, p ij Let x represent the probability that the driver of vehicle j accepts the dispatch request corresponding to dispatch request i; j is the index of the vehicle to be dispatched, 1≤j≤M; i is the index of the dispatch request, 1≤i≤N; x ij This indicates the association information between the vehicle j to be dispatched and the dispatch request i corresponding to the dispatch task; y = 1 indicates that the driver of the vehicle j to be dispatched accepts the dispatch task corresponding to the dispatch request i.

[0172] Step 430: Based on the acceptance probabilities and preset screening rules, determine the target vehicles that match each target grid from all vehicles to be dispatched, wherein the target grid includes at least one dispatching requirement.

[0173] In this embodiment of the application, the above-mentioned screening rules include a scheduling objective function and task distribution constraints, see reference. Figure 7 As shown, when executing step 430, the following steps are specifically performed:

[0174] Step 4301: Based on each acceptance probability, determine the success probability of scheduling the task to be scheduled for each scheduling request.

[0175] In this embodiment of the application, the probability of successful scheduling of the task to be scheduled corresponding to scheduling requirement i can be obtained by the following formula:

[0176]

[0177] Among them, O i p represents the probability of successful scheduling of the task to be scheduled corresponding to scheduling request i; ij This represents the probability that the driver of vehicle j to be dispatched will accept the dispatch task corresponding to dispatch request i.

[0178]

[0179] Step 4302: Based on the scheduling success probability of the task to be scheduled corresponding to each scheduling request, call the preset optimization algorithm to determine each target matching pair so that the above scheduling objective function takes the optimal value under the above task distribution constraints. The above scheduling objective function represents the average scheduling success probability of the task to be scheduled corresponding to all scheduling requests. The task to be scheduled corresponding to the scheduling request in the target matching pair is distributed to the vehicle to be scheduled in the target matching pair.

[0180] In this embodiment of the application, when executing step 4302, the success probability of scheduling each scheduling request can be substituted into the following formula, and a preset optimization algorithm can be called to determine each target matching pair, so that the above scheduling objective function takes the optimal value under the above task distribution constraints:

[0181]

[0182] Where E0 is the scheduling objective function, representing the average success probability of all scheduling requests being accepted by the scheduled tasks; st(subject to) represents the above task distribution constraints: for any scheduled vehicle, it can only receive one scheduled task corresponding to a scheduling request; for any scheduled task corresponding to a scheduling request, it can be pushed to k scheduled vehicles.

[0183] In this embodiment of the application, the preset optimization algorithm can be any heuristic algorithm, such as hill climbing algorithm, genetic algorithm, etc.

[0184] The following section uses the hill-climbing algorithm as an example to detail the specific steps of execution step 4302: Assume there are M vehicles to be scheduled, N scheduling requests, and that the destination of any scheduling request is any POI within the target grid where the corresponding scheduling request is located. Then, the task distribution method can be represented by the following distribution matrix:

[0185]

[0186] Step 1: Based on the task distribution constraints, iterate through all vehicles to be dispatched, find the dispatch request with the highest driver willingness for each vehicle, and match the dispatch request corresponding to the highest driver willingness to the corresponding vehicle. That is, assign the corresponding task in the matrix above to the vehicle. ij Let it be 1. For example, if the scheduling request with the highest driver willingness for vehicle 1 to be scheduled is scheduling request 3, then a in the above distribution matrix is... 31 =1.

[0187] Thus, Step 1 is equivalent to setting an initial value for each element in the above distribution matrix, and obtaining the initial allocation result.

[0188] Step 2: Iterate through all scheduling requests. Based on the matching results of Step 1, calculate the value of the scheduling objective function E0, denoted as E0', for each scheduling request and the acceptance probability of the task to be scheduled.

[0189] Step 3: Iterate through all scheduling requests. For any scheduling request (hereinafter referred to as scheduling request A), based on the task distribution constraints, match each vehicle to be scheduled that was not matched with scheduling request A in Step 1 with scheduling request A. Based on the acceptance probability of the task to be scheduled corresponding to each scheduling request, calculate the value of the new scheduling objective function E0 after scheduling request A is matched with the new vehicle to be scheduled, denoted as E0″. Then, compare E0″ and E0'. If E0″ > E0', retain the matching result of scheduling request A with the new vehicle to be scheduled and cancel the original matching result; otherwise, if E0″ < E0', retain the matching result of scheduling request A with the original vehicle to be scheduled.

[0190] This process continues until all scheduling requests have been traversed, resulting in the final matching result between each scheduling request and the corresponding vehicle to be scheduled (i.e., obtaining the target matching pairs mentioned above).

[0191] Step 4303: Determine the vehicles to be dispatched in each of the above target matching pairs as target dispatch vehicles that match the target grid where the corresponding dispatching requirement is located.

[0192] In this embodiment of the application, after determining each target dispatch vehicle in step 4303, the tasks to be dispatched corresponding to the dispatch requirements in each target matching pair are pushed to the vehicles to be dispatched in the target matching pair (i.e., target dispatch vehicles). In this way, by setting the dispatch objective function and task distribution constraints, each dispatch task distribution combination (i.e., each target matching pair) that maximizes the average dispatch success probability of all dispatch tasks corresponding to the dispatch requirements can be determined from multiple vehicles to be dispatched. Based on each dispatch task distribution combination, the tasks to be dispatched corresponding to all dispatch requirements are distributed, which can improve the overall dispatch success rate of the platform server.

[0193] Based on the same inventive concept, see [reference] Figure 8 As shown in the figure, this application provides a vehicle dispatching device, including:

[0194] The feature information prediction module 810 is used to obtain vehicle feature information in each grid based on the prediction information of each grid within a preset time period in the future;

[0195] The acquisition module 820 is used to determine the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information in each grid, and to acquire the association information of each vehicle to be dispatched.

[0196] The probability prediction module 830 is used to predict the probability of each driver of each vehicle accepting each task based on the association information of each vehicle to be dispatched.

[0197] The determination module 840 is used to determine the target dispatch vehicle that matches each target grid from all vehicles to be dispatched based on each acceptance probability and a preset filtering rule, wherein the target grid includes at least one dispatch requirement.

[0198] In one possible implementation, the acquisition module 820 is used for:

[0199] For any grid in each grid, based on the number of redundant vehicles included in the vehicle feature information in any grid, each target vehicle is determined from the vehicles currently located in any grid, wherein the target vehicle includes at least one of empty vehicles and passenger vehicles whose remaining journey time is less than a preset time.

[0200] If the total number of all target vehicles is not greater than the number of redundant vehicles, then each target vehicle is determined as a vehicle to be dispatched.

[0201] If the total number of all target vehicles is greater than the number of redundant vehicles, then the target vehicles are sorted, and the vehicles to be dispatched are determined from the target vehicles based on the sorting results.

[0202] In one possible implementation, after determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid, and before obtaining the association information of each vehicle to be dispatched, the acquisition module 820 is further configured to:

[0203] Based on the predicted information of each grid within a preset time period in the future, the demand characteristic information within each target grid is obtained;

[0204] Based on the demand feature information within each target grid and each vehicle to be dispatched, a task to be dispatched corresponding to each dispatch demand is generated, wherein the destination of the task to be dispatched is any point of interest (POI) in the target grid where the corresponding dispatch demand is located.

[0205] In one possible implementation, the associated information includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, and real-time characteristics of the driver of the vehicle to be dispatched; and the grid characteristics of the grid where the vehicle to be dispatched is located.

[0206] The acquisition module 820 is specifically used to: acquire the professional characteristics information of the driver of each vehicle to be dispatched; and / or acquire the historical behavioral characteristics information of the driver of each vehicle to be dispatched; and / or acquire the real-time characteristics information of the driver of each vehicle to be dispatched; and / or acquire the grid characteristics information of the grid where each vehicle to be dispatched is located.

[0207] In one possible implementation, the historical behavioral characteristics of the driver of any vehicle to be dispatched are obtained in the following way:

[0208] The first time interval is obtained based on the last dispatch task push time and the current time of any of the vehicles to be dispatched;

[0209] The second time interval is obtained based on the last time the dispatch task was rejected and the current time for any of the vehicles to be dispatched;

[0210] The task acceptance success rate of any vehicle to be dispatched, the average task acceptance time and the average task rejection time of the driver of any vehicle to be dispatched are statistically analyzed within the first preset historical time period.

[0211] The first time interval, the second time interval, the task acceptance success rate, the average task acceptance time, and the average task rejection time are used as part or all of the historical behavioral feature information of the driver of any vehicle to be dispatched.

[0212] In one possible implementation, the real-time characteristic information of the driver of any vehicle to be dispatched is obtained in the following way:

[0213] Based on the current time, determine the time identifier of the task to be scheduled corresponding to any scheduling request;

[0214] Based on the current location of any vehicle to be dispatched and the destination of the task to be dispatched corresponding to any dispatch request, the estimated arrival time is obtained.

[0215] Based on the target time and the current time of any vehicle to be dispatched, the empty driving time is obtained, wherein the target time includes at least one of the last completed order time and the online time of the day;

[0216] Based on the online time of any vehicle to be dispatched on the same day and the current time, the online duration on the same day is obtained, and the number of completed orders of any vehicle to be dispatched within the second preset historical duration is counted.

[0217] The time identifier, the empty driving time, the estimated arrival time, the online time of the day, and the number of completed orders are used as some or all of the real-time characteristic information of the driver of any vehicle to be dispatched.

[0218] In one possible implementation, the grid feature information of any grid containing a vehicle to be scheduled is obtained as follows:

[0219] The acceptance rate, rejection rate, and cancellation rate of the target scheduling task in the grid where any vehicle to be scheduled is located within the third preset historical time period are statistically analyzed. The starting grid of the target scheduling task is the grid where any vehicle to be scheduled is located, and the destination grid is the target grid where the corresponding scheduling request is located.

[0220] Obtain the reference grid information of the grid where the driver of any vehicle to be dispatched is located when the driver accepts the target dispatch task, and the destination grid information of the destination grid;

[0221] The acceptance rate, rejection rate, cancellation rate, reference grid information, and destination grid information are used as part or all of the grid feature information of the grid where any vehicle to be dispatched is located.

[0222] In one possible implementation, the probability prediction module 830 is used for:

[0223] The associated information of each vehicle to be dispatched is input into the target prediction model to obtain the probability of each driver of each vehicle accepting each dispatched task.

[0224] The target prediction model is trained based on a prediction training set, which includes multiple training samples. Each training sample includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, real-time characteristics of any historical scheduling task, and grid characteristics of the vehicle corresponding to the training sample, as well as the driver's acceptance information for any historical scheduling task.

[0225] In one possible implementation, the filtering rules include a scheduling objective function and task distribution constraints; the determining module 840 is used for:

[0226] Based on the aforementioned acceptance probabilities, the success probability of scheduling the task to be scheduled corresponding to each scheduling request is determined.

[0227] Based on the scheduling success probability of the task to be scheduled corresponding to each scheduling request, a preset optimization algorithm is invoked to determine each target matching pair so that the scheduling objective function takes the optimal value under the task distribution constraint. The scheduling objective function represents the average scheduling success probability of the task to be scheduled corresponding to all scheduling requests. The task to be scheduled corresponding to the scheduling request in the target matching pair is distributed to the vehicle to be scheduled in the target matching pair.

[0228] The vehicles to be scheduled in each target matching pair are respectively determined as target scheduling vehicles that match the target grid where the corresponding scheduling requirement is located.

[0229] Based on the same inventive concept, this application provides an electronic device, which is described below with reference to... Figure 9 To describe an electronic device 90 according to this embodiment of the present application. Figure 9 The electronic device 90 shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0230] like Figure 9 As shown, the electronic device 90 is presented in the form of a general-purpose electronic device. The components of the electronic device 90 may include, but are not limited to: at least one processor 91, at least one memory 92, and a bus 93 connecting different system components (including memory 92 and processor 91).

[0231] Bus 93 represents one or more of several bus structures, including a memory bus or memory controller, peripheral bus, processor, or local bus using any of the various bus structures.

[0232] The memory 92 may include a readable medium in the form of volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0233] The memory 92 may also include a program / utility 925 having a set (at least one) of program modules 924, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0234] Electronic device 90 can also communicate with one or more external devices 94 (e.g., keyboard, pointing device, etc.), and with one or more devices that enable a user to interact with electronic device 90, and / or with any device that enables electronic device 90 to communicate with one or more other electronic devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 95. Furthermore, electronic device 90 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 96. As shown, network adapter 96 communicates with other modules used in electronic device 90 via bus 93. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 90, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0235] Based on the same inventive concept, embodiments of this application provide a computer-readable storage medium that, when executed by a processor, enables the processor to perform the method described in any of the preceding claims. Since the principle by which the computer-readable storage medium solves the problem is similar to that of the vehicle scheduling method, the implementation of the computer-readable storage medium can be found in the implementation of the method, and repeated details will not be elaborated further.

[0236] Based on the same inventive concept, this application also provides a computer program product, which includes computer program code. When the computer program code is run on a computer, it causes the computer to execute any of the methods described above. Since the principle by which the above computer program product solves the problem is similar to the method for vehicle dispatching, the implementation of the above computer program product can be referred to the implementation of the method, and repeated details will not be repeated.

[0237] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0238] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0239] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more processes in a flowchart and / or one or more blocks in a block diagram.

[0240] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more blocks in the block diagram.

[0241] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for vehicle dispatching, characterized in that, include: Based on the predicted information of each grid within a preset time period in the future, the vehicle feature information of each grid is obtained; Based on the vehicle feature information within each grid, vehicles to be dispatched are determined from the vehicles currently located in the corresponding grid, and the associated information of each vehicle to be dispatched is obtained. The associated information includes the professional feature information, historical behavior feature information, and real-time feature information of the driver of the vehicle to be dispatched, as well as the grid feature information of the grid where the vehicle to be dispatched is located. Based on the associated information of each vehicle to be dispatched, predict the probability that the driver of each vehicle to be dispatched will accept each dispatched task. Based on the acceptance probabilities and preset screening rules, target vehicles that match each target grid are determined from all vehicles to be dispatched, and the target grid includes at least one dispatching requirement. The real-time characteristic information of the driver of any vehicle to be dispatched includes some or all of the following: a time identifier representing the time period in which the dispatched task corresponding to any dispatch request is located; the estimated arrival time of the vehicle to be dispatched to the destination corresponding to the dispatch request; the empty driving time and the online time of the vehicle to be dispatched on the same day; and the number of completed orders of the vehicle to be dispatched within the second preset historical time period. The grid feature information of any grid where a vehicle to be dispatched is located includes some or all of the following: the acceptance rate, rejection rate, and cancellation rate of the target dispatch task of the grid where the vehicle to be dispatched is located within a third preset historical time period, wherein the starting grid of the target dispatch task is the grid where the vehicle to be dispatched is located, and the destination grid is the target grid where the corresponding dispatch request is located; the reference grid information of the grid where the driver of any vehicle to be dispatched is located when accepting the target dispatch task and the destination grid information of the destination grid.

2. The method as described in claim 1, characterized in that, The step of determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid includes: For any grid in the grid, based on the number of redundant vehicles included in the vehicle feature information in the grid, each target vehicle is determined from the vehicles currently located in the grid. The target vehicles include at least one of empty vehicles and passenger vehicles whose remaining journey time is less than a preset time. If the total number of all target vehicles is not greater than the number of redundant vehicles, then each target vehicle is determined as a vehicle to be dispatched. If the total number of all target vehicles is greater than the number of redundant vehicles, then the target vehicles are sorted, and the vehicles to be dispatched are determined from the target vehicles based on the sorting results.

3. The method as described in claim 1, characterized in that, After determining the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information within each grid, and before obtaining the association information of each vehicle to be dispatched, the process further includes: Based on the predicted information of each grid within a preset time period in the future, the demand characteristic information within each target grid is obtained; Based on the demand feature information within each target grid and each vehicle to be dispatched, a task to be dispatched corresponding to each dispatch demand is generated, wherein the destination of the task to be dispatched is any point of interest (POI) in the target grid where the corresponding dispatch demand is located.

4. The method as described in claim 1, characterized in that, The historical behavioral characteristics of the driver of any of the vehicles to be dispatched are obtained in the following manner: The first time interval is obtained based on the last dispatch task push time and the current time of any of the vehicles to be dispatched; The second time interval is obtained based on the last time the dispatch task was rejected and the current time for any of the vehicles to be dispatched; The task acceptance success rate of any vehicle to be dispatched, the average task acceptance time and the average task rejection time of the driver of any vehicle to be dispatched are statistically analyzed within the first preset historical time period. The first time interval, the second time interval, the task acceptance success rate, the average task acceptance time, and the average task rejection time are used as part or all of the historical behavioral feature information of the driver of any vehicle to be dispatched.

5. The method as described in claim 1, characterized in that, The real-time characteristic information of the driver of any of the vehicles to be dispatched is obtained in the following manner: Based on the current time, determine the time identifier of the task to be scheduled corresponding to any of the scheduling requests; The estimated arrival time is obtained based on the current location of any vehicle to be dispatched and the destination of the task to be dispatched corresponding to any dispatch request. The empty driving time is obtained based on the target time and the current time of any vehicle to be dispatched, wherein the target time includes at least one of the last completed order time and the online time of the day; Based on the online time of any vehicle to be dispatched on the same day and the current time, the online duration on the same day is obtained, and the number of completed orders of any vehicle to be dispatched within the second preset historical duration is counted. The time identifier, the empty driving time, the estimated arrival time, the online time of the day, and the number of completed orders are used as some or all of the real-time characteristic information of the driver of any vehicle to be dispatched.

6. The method as described in claim 1, characterized in that, The grid feature information of the grid where any vehicle to be dispatched is located is obtained in the following manner: The acceptance rate, rejection rate, and cancellation rate of the target scheduling task in the grid where any vehicle to be scheduled is located within the third preset historical time period are statistically obtained. Obtain the reference grid information of the grid where the driver of any vehicle to be dispatched is located when the driver accepts the target dispatch task, and the destination grid information of the destination grid; The acceptance rate, rejection rate, cancellation rate, reference grid information, and destination grid information are used as part or all of the grid feature information of the grid where any vehicle to be dispatched is located.

7. The method according to any one of claims 1-6, characterized in that, The prediction of the driver's acceptance probability for each dispatched task based on the association information of each vehicle to be dispatched includes: Input the associated information of each vehicle to be dispatched into the target prediction model to obtain the probability of each driver of each vehicle accepting each dispatched task. The target prediction model is trained based on a prediction training set, which includes multiple training samples. Each training sample includes at least one of the following: the driver's professional characteristics, historical behavior characteristics, real-time characteristics of any historical scheduling task, and grid characteristics of the vehicle corresponding to the training sample, as well as the driver's acceptance information for any historical scheduling task.

8. The method according to any one of claims 1-6, characterized in that, The screening rules include a scheduling objective function and task distribution constraints; the process of determining the target scheduling vehicle matching each target grid from all vehicles to be scheduled based on each acceptance probability and the preset screening rules includes: Based on the aforementioned acceptance probabilities, the success probability of scheduling the task to be scheduled corresponding to each scheduling request is determined. Based on the scheduling success probability of the task to be scheduled corresponding to each scheduling request, a preset optimization algorithm is invoked to determine each target matching pair so that the scheduling objective function takes the optimal value under the task distribution constraint. The scheduling objective function represents the average scheduling success probability of the task to be scheduled corresponding to all scheduling requests. The task to be scheduled corresponding to the scheduling request in the target matching pair is distributed to the vehicle to be scheduled in the target matching pair. The vehicles to be scheduled in each target matching pair are respectively determined as target scheduling vehicles that match the target grid where the corresponding scheduling requirement is located.

9. A vehicle dispatching device, characterized in that, include: The feature information prediction module is used to obtain vehicle feature information in each grid based on the prediction information of each grid within a preset time period in the future; The acquisition module is used to determine the vehicles to be dispatched from the vehicles currently located in the corresponding grid based on the vehicle feature information in each grid, and to acquire the association information of each vehicle to be dispatched. The association information includes the professional feature information, historical behavior feature information, and real-time feature information of the driver of the corresponding vehicle to be dispatched, as well as the grid feature information of the grid where the corresponding vehicle to be dispatched is located. The probability prediction module is used to predict the probability that the driver of each vehicle to be dispatched will accept each dispatched task based on the association information of each vehicle to be dispatched. The determination module is used to determine the target dispatch vehicle that matches each target grid from all vehicles to be dispatched based on each acceptance probability and preset filtering rules, wherein the target grid includes at least one dispatch requirement; The real-time characteristic information of the driver of any vehicle to be dispatched includes some or all of the following: a time identifier representing the time period in which the dispatched task corresponding to any dispatch request is located; the estimated arrival time of the vehicle to be dispatched to the destination corresponding to the dispatch request; the empty driving time and the online time of the vehicle to be dispatched on the same day; and the number of completed orders of the vehicle to be dispatched within the second preset historical time period. The grid feature information of any grid where a vehicle to be dispatched is located includes some or all of the following: the acceptance rate, rejection rate, and cancellation rate of the target dispatch task of the grid where the vehicle to be dispatched is located within a third preset historical time period, wherein the starting grid of the target dispatch task is the grid where the vehicle to be dispatched is located, and the destination grid is the target grid where the corresponding dispatch request is located; the reference grid information of the grid where the driver of any vehicle to be dispatched is located when accepting the target dispatch task and the destination grid information of the destination grid.

10. An electronic device, characterized in that, include: Memory is used to store computer programs or instructions; A processor for executing a computer program or instructions in the memory such that the method described in any one of claims 1-8 is performed.

11. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor, the processor is able to perform the method as described in any one of claims 1-8.

12. A computer program product, characterized in that, The computer program product includes: computer program code, which, when run on a computer, causes the computer to perform the method described in any one of claims 1-8.

Images