A vehicle scheduling method, device, electronic device and readable storage medium

Abstract

The present application provides a vehicle scheduling method, including: obtaining passenger characteristics within a scheduling area and vehicle characteristics corresponding to each target vehicle; inputting the passenger characteristics within the scheduling area into a regional vehicle usage expectation prediction model for vehicle usage expectation prediction processing to obtain the vehicle usage expectation information of the scheduling area; fusing the vehicle characteristics corresponding to each target vehicle within the scheduling area with the vehicle usage expectation information corresponding to the scheduling area respectively to obtain the fusion characteristics corresponding to each target vehicle within the scheduling area; inputting the fusion characteristics corresponding to each target vehicle within the scheduling area into a scheduling area prediction model for scheduling area prediction processing to obtain the predicted scheduling areas of each target vehicle within the scheduling area; scheduling the target vehicles in the multiple scheduling areas according to the predicted scheduling areas of each target vehicle in each scheduling area; and using a dual intelligent model to accurately predict the distribution of passenger vehicle usage demands and the vehicle scheduling target areas.

Description

Technical Field

[0001] This application relates to the field of intelligent technologies, and particularly to a vehicle scheduling method, device, electronic device, and readable storage medium. Background Art

[0002] To date, in the taxi-hailing industry, there are still a large number of problems caused by the mismatch and irrationality of the demands between the client side and the driver side, such as difficult taxi-hailing for users or long waiting times, and long pick-up distances for drivers.

[0003] In the existing technology, the deep reinforcement learning technology for operating vehicle resource scheduling mainly uses the methods and systems of single-agent deep reinforcement learning. Due to the complexity of the taxi market and the user market, the methods and systems of single-agent deep reinforcement learning cannot accurately predict the distribution of passenger car usage demands and the target areas for vehicle scheduling. Summary of the Invention

[0004] This application provides a vehicle scheduling method, device, electronic device, and readable storage medium that can accurately predict the distribution of passenger car usage demands and the target areas for vehicle scheduling to overcome the above-mentioned defects in the existing technology.

[0005] To solve the above technical problems, this application provides the following technical solutions:

[0006] According to the first aspect of the embodiments of this application, a vehicle scheduling method is provided, including:

[0007] Obtain the passenger characteristics within the scheduling area and the vehicle characteristics corresponding to each target vehicle; the scheduling area is any one of a plurality of scheduling areas obtained by dividing a preset geographical range, and the target vehicle is an operating vehicle in an idle state;

[0008] Input the passenger characteristics within the scheduling area into the regional car usage expectation prediction model for car usage expectation prediction processing to obtain the car usage expectation information of the scheduling area;

[0009] Fuse the vehicle characteristics corresponding to each target vehicle within the scheduling area with the car usage expectation information corresponding to the scheduling area to obtain the fusion characteristics corresponding to each target vehicle within the scheduling area;

[0010] Input the fusion characteristics corresponding to each target vehicle within the scheduling area into the scheduling area prediction model for scheduling area prediction processing to obtain the predicted scheduling areas of each target vehicle within the scheduling area;

[0011] Schedule the target vehicles in the plurality of scheduling areas according to the predicted scheduling areas of each target vehicle in each scheduling area.

[0012] In an exemplary embodiment, the method further includes:

[0013] Obtain the operation status information of each operating vehicle in the scheduling area;

[0014] Determine the operating vehicles in the idle state according to the operation status information to obtain the target vehicles in the scheduling area.

[0015] In an exemplary embodiment, the method further includes:

[0016] Obtain the sample passenger characteristics of the sample scheduling area at a first moment, and the vehicle usage label information corresponding to the sample scheduling area; the vehicle usage label information is determined according to the reservation vehicle usage information and the actual vehicle usage information of the sample scheduling area at a second moment, and the second moment refers to a moment after the first moment;

[0017] Input the sample passenger characteristics of the sample scheduling area into a first preset neural network model for vehicle usage expectation prediction to obtain the vehicle usage expectation information corresponding to the sample scheduling area;

[0018] Adjust the model parameters of the first preset neural network model according to the difference between the vehicle usage expectation information corresponding to the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area;

[0019] Continue iterative training based on the adjusted model parameters until a preset training end condition is reached to obtain the regional vehicle usage expectation prediction model.

[0020] In an exemplary embodiment, the method further includes:

[0021] Obtain the reservation vehicle usage information and the actual vehicle usage information of the sample scheduling area at the second moment;

[0022] Determine the weights corresponding to the reservation vehicle usage information and the actual vehicle usage information respectively;

[0023] Perform weighted summation according to the reservation vehicle usage information, the actual vehicle usage information and the corresponding weights to obtain the vehicle usage label information.

[0024] In an exemplary embodiment, the method further includes:

[0025] Obtain the sample vehicle characteristics of each sample vehicle in the sample scheduling area at a third moment, and the scheduling label information corresponding to the sample scheduling area; the scheduling label information is determined according to the order receiving information of the dispatched vehicles in the sample scheduling area at a fourth moment, and the third moment refers to a moment after the fourth moment;

[0026] Fuse the sample vehicle features of each of the sample vehicles with the vehicle usage expectation information corresponding to the sample scheduling area to obtain the fusion features corresponding to each of the sample vehicles, and input the fusion features corresponding to each of the sample vehicles into a second preset neural network model for scheduling area prediction to obtain the scheduling area prediction results corresponding to each of the sample vehicles;

[0027] Adjust the model parameters of the second preset neural network model according to the differences between the scheduling area prediction results corresponding to each of the sample vehicles and the scheduling label information;

[0028] Continue iterative training based on the adjusted model parameters until a preset training end condition is reached to obtain the scheduling area prediction model.

[0029] In an exemplary embodiment, the method further includes:

[0030] Obtain the order receiving information of the vehicles receiving orders in the sample scheduling area at the fourth moment, where the order receiving information includes the order receiving volume and the vehicle usage settlement information;

[0031] Determine the weights corresponding to the order receiving volume and the vehicle usage settlement information respectively;

[0032] Perform weighted summation according to the order receiving volume, the vehicle usage settlement information, and the corresponding weights to obtain the scheduling label information.

[0033] In an exemplary embodiment, the first preset neural network model and the second preset neural network model are deep neural networks or multi-layer feedforward neural networks based on the error backpropagation algorithm.

[0034] According to the second aspect of the embodiments of the present application, there is provided a vehicle scheduling device, and the device includes:

[0035] A feature acquisition module, configured to acquire passenger features in a scheduling area and vehicle features corresponding to each target vehicle, where the scheduling area is any one of a plurality of areas obtained by dividing a preset geographical range, and the target vehicle is an operating vehicle in an idle state;

[0036] A regional vehicle usage expectation prediction module, configured to acquire the vehicle usage expectation information of the scheduling area based on the passenger features in the scheduling area;

[0037] A vehicle usage expectation information extraction module, configured to extract the vehicle usage expectation information of the scheduling area;

[0038] A feature fusion module, configured to acquire the fusion features corresponding to each target vehicle in the scheduling area, where the fusion features corresponding to each target vehicle in the scheduling area are based on the vehicle features of each target vehicle in the scheduling area and the vehicle usage expectation information corresponding to the scheduling area;

[0039] A scheduling area prediction module, configured to obtain a predicted scheduling area of each target vehicle in the scheduling area based on the fusion features corresponding to each target vehicle in the scheduling area;

[0040] A target vehicle scheduling module, configured to schedule the target vehicles in the multiple scheduling areas based on the predicted scheduling areas of the target vehicles in the scheduling area.

[0041] In an exemplary embodiment, the apparatus further includes:

[0042] A feature acquisition module, configured to obtain target vehicles in the scheduling area based on the operation status information of each operating vehicle in the scheduling area.

[0043] In an exemplary embodiment, the apparatus further includes:

[0044] A first sample feature acquisition module, configured to obtain sample passenger features of a sample scheduling area at a first moment, and vehicle use label information corresponding to the sample scheduling area; the vehicle use label information is determined according to reservation vehicle use information and actual vehicle use information of the sample scheduling area at a second moment, and the second moment refers to a moment after the first moment;

[0045] A sample area vehicle use expectation prediction module, configured to obtain vehicle use expectation information corresponding to the sample scheduling area based on the sample passenger features of the sample scheduling area and a first preset neural network module;

[0046] A first sample parameter adjustment module, configured to adjust module parameters of the first preset neural network module based on the difference between the vehicle use expectation information corresponding to the sample scheduling area and the vehicle use label information corresponding to the sample scheduling area;

[0047] A first judgment module, configured to judge whether the module parameters of the adjusted first preset neural network module reach a preset training end condition according to the difference between the vehicle use expectation information corresponding to the sample scheduling area and the vehicle use label information corresponding to the sample scheduling area.

[0048] In an exemplary embodiment, the apparatus further includes:

[0049] A second sample feature acquisition module, configured to obtain sample vehicle features of each sample vehicle in the sample scheduling area at a third moment, and scheduling label information corresponding to the sample scheduling area; the scheduling label information is determined according to the order receiving information of the dispatched vehicles in the sample scheduling area at a fourth moment, and the third moment refers to a moment after the fourth moment;

[0050] A sample vehicle use expectation information extraction module, configured to extract the vehicle use expectation information corresponding to the sample scheduling area from the sample area vehicle use expectation prediction module;

[0051] A sample feature fusion module, configured to obtain the fusion features corresponding to each sample vehicle, where the fusion features corresponding to each sample vehicle are based on the sample vehicle features of each sample vehicle and the vehicle usage expectation information corresponding to the sample scheduling area;

[0052] A sample scheduling area prediction module, configured to obtain the scheduling area prediction results corresponding to each sample vehicle based on the fusion features corresponding to each sample vehicle and a second preset neural network module;

[0053] A second sample parameter adjustment module, configured to adjust the module parameters of the second preset neural network module based on the difference between the scheduling area prediction results corresponding to each sample vehicle and the scheduling label information;

[0054] A second judgment module, configured to compare the difference between the scheduling area prediction results corresponding to each sample vehicle and the scheduling label information, and adjust the module parameters of the second preset neural network module according to the difference.

[0055] According to the third aspect of the embodiments of the present application, there is provided an electronic device, including a processor and a memory, where at least one instruction or at least one program segment is stored in the memory, and the at least one instruction or the at least one program segment is loaded and executed by the processor to implement the above vehicle scheduling method.

[0056] According to the fourth aspect of the embodiments of the present application, there is provided a computer-readable storage medium, where at least one instruction or at least one program segment is stored in the storage medium, and the at least one instruction or the at least one program segment is loaded and executed by a processor to implement the above vehicle scheduling method.

[0057] By adopting the above technical solutions, the present application has the following beneficial effects:

[0058] A vehicle scheduling method, device, electronic device and readable storage medium provided by the present application adopt a technical solution of a dual intelligent model of a regional vehicle usage expectation prediction model and a scheduling area prediction model, and both the regional vehicle usage expectation prediction model and the scheduling area prediction model are real-time updated deep reinforcement learning systems, which are more suitable for simulating and predicting various uncertain factors in complex scenarios, making the prediction results closer to the real vehicle usage scenarios, so as to achieve the effect of accurately predicting the distribution of passenger vehicle usage demands and the vehicle scheduling target areas, promoting the distribution of vehicle resources to be more in line with the user's demand expectations for vehicle resources, and realizing reasonable resource scheduling. Description of the Drawings

[0059] To more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other accompanying drawings can be obtained based on these drawings without creative efforts.

[0060] Figure 1 A flowchart of a vehicle scheduling method provided by an embodiment of the present application;

[0061] Figure 2 A training flowchart of a vehicle usage expectation prediction model provided by an embodiment of the present application;

[0062] Figure 3 A training flowchart of a scheduling area prediction model provided by an embodiment of the present application;

[0063] Figure 4 A structural block diagram of a vehicle scheduling device provided by an embodiment of the present application;

[0064] Figure 5 A hardware structural block diagram of an electronic device for running a vehicle scheduling method provided by an embodiment of the present application. Detailed implementation manners

[0065] The following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts fall within the protection scope of the present application.

[0066] As used herein, "one embodiment" or "an embodiment" refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present application. In the description of the embodiments of the present application, it should be understood that the orientation or positional relationship indicated by terms such as "upper", "lower", "top", "bottom", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present application. In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. Moreover, the terms "first", "second", etc. are used to distinguish similar objects and do not necessarily have to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein.

[0067] Please refer to Figure 1 , which shows a schematic flow chart of a vehicle scheduling method provided by an embodiment of the present application. The vehicle scheduling method includes:

[0068] Step S101: Obtain the passenger characteristics within the scheduling area and the vehicle characteristics corresponding to each target vehicle; the scheduling area is any one of multiple scheduling areas obtained by dividing a preset geographical range, and the target vehicle is an operating vehicle in an idle state;

[0069] Step S102: Input the passenger characteristics within the scheduling area into the regional vehicle usage expectation prediction model for vehicle usage expectation prediction processing to obtain the vehicle usage expectation information of the scheduling area;

[0070] Step S103: Integrate the vehicle characteristics corresponding to each target vehicle within the scheduling area with the vehicle usage expectation information corresponding to the scheduling area to obtain the integrated characteristics corresponding to each target vehicle within the scheduling area;

[0071] Step S104: Input the integrated characteristics corresponding to each target vehicle within the scheduling area into the scheduling area prediction model for scheduling area prediction processing to obtain the predicted scheduling areas of each target vehicle within the scheduling area;

[0072] Step S105: Schedule the target vehicles in multiple scheduling areas according to the predicted scheduling areas of each target vehicle in each scheduling area.

[0073] In a specific embodiment, a preset geographical range is pre-divided to obtain multiple scheduling areas; the passenger characteristics and the operation status information of each operating vehicle within the scheduling area are obtained through step S101. Among them, through the operation status information of each operating vehicle within the scheduling area, the operating vehicles in the idle state, that is, the target vehicles, are determined, and then the vehicle characteristics corresponding to each target vehicle are extracted; the vehicle usage expectation information of the scheduling area, that is, the vehicle usage demand, is predicted through step S102. Among them, in order to make the predicted vehicle usage expectation information of the scheduling area closer to the real vehicle usage scenario and achieve the purpose of accurately predicting the distribution of passenger demands, the deep reinforcement learning system of the regional vehicle usage expectation prediction model is an online learning system that is updated in real time; the fusion characteristics corresponding to each target vehicle within the scheduling area are obtained through step S103; the predicted scheduling area of each target vehicle is predicted through step S104. Among them, in order to make the predicted scheduling area fit the actual demand and achieve the effect of accurately predicting the target area of vehicle scheduling, the deep reinforcement learning system of the scheduling area prediction model is an online learning system that is updated in real time; an instruction is sent to the target vehicle through step S105 to make the target vehicle go to the predicted scheduling area, so as to achieve the purpose of vehicle scheduling. In this embodiment, the technical solution of a dual-intelligent model of a regional vehicle usage expectation prediction model and a scheduling area prediction model is adopted, and both the regional vehicle usage expectation prediction model and the scheduling area prediction model are deep reinforcement learning systems that are updated in real time, which are more suitable for simulating and predicting various uncertain factors in complex scenarios, making the prediction results closer to the real vehicle usage scenario, so as to achieve the effect of accurately predicting the distribution of passenger vehicle usage demands and the target area of vehicle scheduling, promoting the distribution of vehicle resources to better meet the user's demand expectations for vehicle resources, and realizing reasonable resource scheduling.

[0074] It should be noted that the passenger characteristics within the scheduling area and the vehicle characteristics corresponding to each target vehicle include the real-time positioning coordinates of the passengers and the target vehicles, the historical taxi-taking habits of the passengers, the real-time traffic congestion situation, the date characteristics of the day, the real-time weather conditions of the areas where the passengers and the target vehicles are located, and the weather conditions at the next moment.

[0075] Specifically, the embodiment of the present application may further include the step of determining the target vehicle, and this step specifically includes:

[0076] Obtain the operation status information of each operating vehicle within the scheduling area;

[0077] According to the operation status information, determine the operating vehicles in the idle state to obtain the target vehicles within the scheduling area.

[0078] In a specific embodiment, first, determine the operation vehicles in the idle state through the operation state information of each operation vehicle, and then define the operation vehicles in the idle state as the target vehicles in the dispatching area, which can avoid the situation of wrong dispatching of the operation vehicles in the carrying state during the dispatching process and causing passengers to wait in vain.

[0079] Please refer to Figure 2 , which shows a training flow chart of a vehicle usage expectation prediction model provided by an embodiment of the present application. The training of the vehicle usage expectation prediction model includes:

[0080] Step S201: Obtain the sample passenger characteristics of the sample dispatching area at the first moment, and the vehicle usage label information corresponding to the sample dispatching area; the vehicle usage label information is determined according to the reserved vehicle usage information and the actual vehicle usage information of the sample dispatching area at the second moment, and the second moment refers to the moment after the first moment;

[0081] Step S202: Input the sample passenger characteristics of the sample dispatching area into the first preset neural network model for vehicle usage expectation prediction to obtain the vehicle usage expectation information corresponding to the sample dispatching area;

[0082] Step S203: Adjust the model parameters of the first preset neural network model according to the difference between the vehicle usage expectation information corresponding to the sample dispatching area and the vehicle usage label information corresponding to the sample dispatching area;

[0083] Step S204: Determine whether the model parameters of the adjusted first preset neural network model reach the preset training end condition;

[0084] If not, continue iterative training based on the adjusted model parameters;

[0085] If so, execute Step S205: Obtain the regional vehicle usage expectation prediction model.

[0086] In a specific embodiment, sample passenger characteristics at each historical moment within a sample scheduling area are collected in advance, as well as the vehicle usage label information corresponding to the sample scheduling area, that is, the reservation vehicle usage information and the actual vehicle usage information at each historical moment; the sample passenger characteristics at the first moment in the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area are obtained through step S201, that is, the reservation vehicle usage information and the actual vehicle usage information at the second moment; the vehicle usage expectation information corresponding to the sample scheduling area is obtained through step S202, that is, the vehicle usage demand corresponding to the sample scheduling area; the difference between the vehicle usage expectation information corresponding to the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area is compared through step S203, and the model parameters of the first preset neural network model are adjusted according to this difference, where the first preset neural network model is a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm; it is judged through step S204 whether the model parameters of the first preset neural network model after being adjusted through step S203 reach the preset training end condition. If the preset training end condition is not reached, based on the model parameters of the adjusted first preset neural network model, return to step S201 for iterative training until the model parameters of the first preset neural network model after being adjusted through step S203 reach the preset training end condition and meet the determination condition of step S204, then execute step S205; through step S205, the regional vehicle usage expectation prediction model required by step S102 is obtained. This embodiment is based on the collected historical data, and through steps S201 to S205, a regional vehicle usage expectation prediction model that meets the preset conditions is trained. This regional vehicle usage expectation prediction model will be updated in real time during actual application. Through the online learning method, it can accurately predict the distribution of passenger vehicle usage demands, promote the distribution of vehicle resources to better meet the user's demand expectations for vehicle resources, and achieve reasonable resource scheduling.

[0087] It should be noted that the training of the regional vehicle usage expectation prediction model is an iterative training based on historical data. Only when the model parameters of the first preset neural network model reach the preset training end condition can the scheduling area prediction model required for executing step S102 be obtained.

[0088] It should also be noted that during the iterative training process, each time step S201 is executed, the first moment and the second moment obtained are both subsequent moments based on the first moment and the second moment obtained when step S201 was executed last time.

[0089] It should also be noted that the reservation vehicle usage information can specifically be the number of reservation vehicle usage orders, and the actual vehicle usage information can specifically be the number of actual vehicle usage completed orders.

[0090] Specifically, the embodiment of the present application may further include a step of determining the above vehicle usage label information, and this step specifically includes:

[0091] Obtain the reserved car - using information and actual car - using information of the sample scheduling area at the second moment;

[0092] Determine the weights corresponding to the reserved car - using information and the actual car - using information respectively;

[0093] Perform weighted summation according to the reserved car - using information, the actual car - using information and the corresponding weights to obtain the car - using label information.

[0094] In a specific embodiment, the car - using label information is obtained by weighted summation of the reserved car - using information, the actual car - using information and the corresponding weights. Among them, the reserved car - using information can be specifically the number of reserved car - using orders, and the actual car - using information can be specifically the number of actual car - using completed orders. That is, the car - using label information can be obtained by weighted summation of the number of reserved car - using orders, the number of actual car - using completed orders and the corresponding weights.

[0095] Please refer to Figure 3 , which shows a training flow chart of a scheduling area prediction model provided by an embodiment of the present application. The training of the scheduling area prediction model includes:

[0096] Step S301: Obtain the sample vehicle characteristics of each sample vehicle in the sample scheduling area at the third moment, and the scheduling label information corresponding to the sample scheduling area; the scheduling label information is determined according to the order - receiving information of the dispatched vehicles in the sample scheduling area at the fourth moment, and the third moment refers to the moment after the fourth moment;

[0097] Step S302: Integrate the sample vehicle characteristics of each sample vehicle and the car - using expectation information corresponding to the sample scheduling area to obtain the integrated characteristics corresponding to each sample vehicle;

[0098] Step S303: Input the integrated characteristics corresponding to each sample vehicle into the second preset neural network model for scheduling area prediction to obtain the scheduling area prediction results corresponding to each sample vehicle;

[0099] Step S304: Adjust the model parameters of the second preset neural network model according to the difference between the scheduling area prediction results corresponding to each sample vehicle and the scheduling label information;

[0100] Step S305: Determine whether the model parameters of the adjusted second preset neural network model reach the preset training end condition;

[0101] If not, continue iterative training based on the adjusted model parameters;

[0102] If so, then S306: Obtain the scheduling area prediction model.

[0103] In a specific embodiment, the characteristics of each sample vehicle at each historical moment within the sample scheduling area are collected in advance, as well as the scheduling label information corresponding to the sample scheduling area, that is, the order receiving information of the vehicles to be scheduled at each historical moment; the characteristics of each sample vehicle at the third moment within the sample scheduling area are obtained through step S301, as well as the scheduling label information corresponding to the sample scheduling area, that is, the order receiving information of the vehicles to be scheduled at the second moment; the fusion characteristics corresponding to each sample vehicle are obtained through step S302; the scheduling area prediction results corresponding to each sample vehicle are obtained through step S303, that is, the predicted scheduling areas corresponding to each sample vehicle; the differences between the scheduling area prediction results corresponding to each sample vehicle and the scheduling label information are compared through step S304, and the model parameters of the second preset neural network model are adjusted according to this difference, where the second preset neural network model is a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm; it is judged through step S305 whether the model parameters of the second preset neural network model after being adjusted through step S304 reach the preset training end condition. If the preset training end condition is not reached, based on the model parameters of the adjusted second preset neural network model, return to step S301 for iterative training until the model parameters of the second preset neural network model after being adjusted through step S304 reach the preset training end condition and meet the determination condition of step S305, then execute step S306; through step S306, the scheduling area prediction model required for step S104 is obtained. This embodiment is based on the collected historical data, and a scheduling area prediction model meeting the preset conditions is trained through steps S301 to S306. This scheduling area prediction model will be updated in real time during actual application. Through the method of online learning, it can accurately predict the vehicle scheduling target area, promote the distribution of vehicle resources to better meet the user's demand expectations for vehicle resources, and achieve reasonable resource scheduling.

[0104] It should be noted that the training of the scheduling area prediction model is an iterative training based on historical data. Only when the model parameters of the second preset neural network model reach the preset training end condition can the scheduling area prediction model required for executing step S104 be obtained.

[0105] It should also be noted that during the iterative training process, each time step S301 is executed, the third moment and the fourth moment obtained are both subsequent moments based on the third moment and the fourth moment obtained in the previous execution of step S301.

[0106] It should also be noted that the order receiving information includes the order receiving volume and the vehicle usage settlement information, where the vehicle usage settlement information can specifically be the vehicle usage settlement amount.

[0107] Specifically, the embodiment of the present application may further include the step of determining the above-mentioned scheduling label information, and this step specifically includes:

[0108] Obtain the order-taking information of the vehicles scheduled in the sample scheduling area at the fourth moment, where the order-taking information includes the order-taking volume and the vehicle usage settlement information;

[0109] Determine the weights corresponding to the order-taking volume and the vehicle usage settlement information respectively;

[0110] Perform weighted summation according to the order-taking volume, the vehicle usage settlement information, and the corresponding weights to obtain the scheduling label information.

[0111] In a specific embodiment, the scheduling label information is obtained by weighted summation of the order-taking volume, the vehicle usage settlement information, and the corresponding weights. Among them, the vehicle usage settlement information can be specifically the vehicle usage settlement amount, that is, the scheduling label information is obtained by weighted summation of the order-taking volume, the vehicle usage settlement amount, and the corresponding weights.

[0112] Specifically, the first preset neural network model and the second preset neural network model are deep neural networks or multi-layer feedforward neural networks based on the error backpropagation algorithm.

[0113] In a specific embodiment, using a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm can represent complex functions with fewer parameters and reduce the error signal, which is beneficial to ensuring the stability and accuracy of the operation of the first preset neural network model and the second preset neural network model, and further ensuring the stability and accuracy of the operation of the regional vehicle usage expectation prediction model and the scheduling area prediction model.

[0114] As can be seen from the above technical solutions of the embodiments of the present application, the embodiments of the present application adopt the technical solution of a dual-intelligent model of a regional vehicle usage expectation prediction model and a scheduling area prediction model, and both the regional vehicle usage expectation prediction model and the scheduling area prediction model are deep reinforcement learning systems that are updated in real time, which are more suitable for simulating and predicting various uncertain factors in complex scenarios, making the prediction results closer to the real vehicle usage scenarios, so as to achieve the effect of accurately predicting the distribution of passenger vehicle usage needs and the vehicle scheduling target area, promoting the distribution of vehicle resources to be more in line with the user's demand expectations for vehicle resources, and realizing reasonable resource scheduling.

[0115] Corresponding to the vehicle scheduling method provided in the above embodiment, the embodiment of the present application also provides a vehicle scheduling device. Since the vehicle scheduling device provided in the embodiment of the present application corresponds to the vehicle scheduling method provided in the above embodiment, the implementation manners of the foregoing vehicle scheduling method are also applicable to the vehicle scheduling device provided in this embodiment and will not be described in detail in this embodiment.

[0116] Please refer to Figure 4 , which shows the structural block diagram of a vehicle scheduling device proposed in the embodiment of the present application. The vehicle scheduling device includes:

[0117] A feature acquisition module, configured to acquire passenger features within a scheduling area and vehicle features corresponding to each target vehicle. The scheduling area is any one of multiple areas obtained by dividing a preset geographical range, and the target vehicle is an operating vehicle in an idle state;

[0118] A regional vehicle usage expectation prediction module, configured to obtain vehicle usage expectation information for the scheduling area based on the passenger features within the scheduling area;

[0119] A vehicle usage expectation information extraction module, configured to extract the vehicle usage expectation information for the scheduling area;

[0120] A feature fusion module, configured to obtain the fusion features corresponding to each target vehicle within the scheduling area. The fusion features corresponding to each target vehicle within the scheduling area are based on the vehicle features of each target vehicle within the scheduling area and the vehicle usage expectation information corresponding to the scheduling area;

[0121] A scheduling area prediction module, configured to obtain the predicted scheduling areas of each target vehicle within the scheduling area based on the fusion features corresponding to each target vehicle within the scheduling area;

[0122] A target vehicle scheduling module, configured to schedule the target vehicles in multiple scheduling areas based on the predicted scheduling areas of each target vehicle within the scheduling area.

[0123] In a specific embodiment, a preset geographical range is pre-divided to obtain multiple scheduling areas; a passenger feature acquisition module acquires the passenger features and the operation status information of each operating vehicle within the scheduling area. Among them, based on the operation status information of each operating vehicle within the scheduling area, the operating vehicles in the idle state, that is, the target vehicles, are determined, and then the vehicle features corresponding to each target vehicle are extracted; a regional vehicle usage expectation prediction module acquires the vehicle usage expectation information of the scheduling area, that is, the vehicle usage demand. In order to make the predicted vehicle usage expectation information of the scheduling area closer to the real vehicle usage scenario and achieve the purpose of accurately predicting the distribution of passenger demands, the regional vehicle usage expectation prediction module predicts the vehicle usage expectation information of the scheduling area by means of online learning, using a deep reinforcement learning system related to the regional vehicle usage expectation prediction that is updated in real time; a vehicle usage expectation information extraction module extracts the vehicle usage expectation information of the scheduling area from the regional vehicle usage expectation prediction module; a feature fusion module acquires the fusion features corresponding to each target vehicle within the scheduling area; a scheduling area prediction module acquires the predicted scheduling areas of each target vehicle within the scheduling area. In order to make the predicted scheduling area fit the actual demand and achieve the effect of accurately predicting the target area of vehicle scheduling, the scheduling area prediction module predicts the scheduling area of the target vehicle by means of online learning, using a deep reinforcement learning system related to the scheduling area prediction that is updated in real time; a target vehicle scheduling module sends an instruction to the target vehicle to make the target vehicle go to the predicted scheduling area, achieving the purpose of vehicle scheduling. In this embodiment, by adopting the technical solution of a dual-intelligent module including a regional vehicle usage expectation prediction module and a scheduling area prediction module, and both the regional vehicle usage expectation prediction module and the scheduling area prediction module are deep reinforcement learning systems updated in real time, it is more suitable for simulating and predicting various uncertain factors in complex scenarios, making the prediction result closer to the real vehicle usage scenario, so as to achieve the effect of accurately predicting the distribution of passenger vehicle usage demands and the target area of vehicle scheduling, promoting the distribution of vehicle resources to better meet the user's demand expectation for vehicle resources and realizing reasonable resource scheduling.

[0124] It should be noted that the passenger features within the scheduling area and the vehicle features corresponding to each target vehicle include the real-time positioning coordinates of the passengers and the target vehicles, the historical taxi-taking habits of the passengers, the real-time traffic congestion situation, the date feature of the day, the real-time weather conditions of the areas where the passengers and the target vehicles are located, and the weather conditions at the next moment.

[0125] Specifically, the embodiment of the present application may further include a step of determining the target vehicle, and this step specifically includes:

[0126] A feature acquisition module acquires the target vehicles within the scheduling area based on the operation status information of each operating vehicle within the scheduling area.

[0127] In a specific embodiment, first, the operation status information of each operating vehicle in the scheduling area is obtained through the feature acquisition module, and then, based on the operation status information of each operating vehicle in the scheduling area, the operating vehicles in the idle state are determined, and the operating vehicles in the idle state are designated as the target vehicles in the scheduling area, which can avoid the situation of wrong scheduling of the operating vehicles in the carrying state during the scheduling process, resulting in passengers waiting in vain.

[0128] Specifically, the training process of the regional vehicle usage expectation prediction module may further be included in the embodiments of the present application, and this process specifically includes:

[0129] The first sample feature acquisition module is used to obtain the sample passenger features of the sample scheduling area at the first moment, and the vehicle usage label information corresponding to the sample scheduling area; the vehicle usage label information is determined according to the reservation vehicle usage information and the actual vehicle usage information of the sample scheduling area at the second moment, and the second moment refers to the moment after the first moment;

[0130] The sample area vehicle usage expectation prediction module is used to obtain the vehicle usage expectation information corresponding to the sample scheduling area based on the sample passenger features of the sample scheduling area and the first preset neural network module;

[0131] The first sample parameter adjustment module is used to adjust the module parameters of the first preset neural network module based on the difference between the vehicle usage expectation information corresponding to the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area;

[0132] The first judgment module is used to judge whether the module parameters of the adjusted first preset neural network module reach the preset training end condition according to the difference between the vehicle usage expectation information corresponding to the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area.

[0133] In a specific embodiment, sample passenger characteristics at each historical moment within the sample scheduling area are collected in advance, as well as the vehicle usage label information corresponding to the sample scheduling area, that is, the reservation vehicle usage information and the actual vehicle usage information at each historical moment; the first sample feature acquisition module is used to obtain the sample passenger characteristics of the sample scheduling area at the first moment, as well as the vehicle usage label information corresponding to the sample scheduling area, that is, the reservation vehicle usage information and the actual vehicle usage information at the second moment; the vehicle usage expectation prediction module for the sample area is used to obtain the vehicle usage expectation information corresponding to the sample scheduling area, that is, the vehicle usage demand corresponding to the sample scheduling area; the first sample parameter adjustment module compares the difference between the vehicle usage expectation information corresponding to the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area, and adjusts the module parameters of the first preset neural network module according to this difference, where the first preset neural network module is a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm; the first judgment module judges whether the module parameters of the first preset neural network module adjusted by the first sample parameter adjustment module reach the preset training end condition. If the preset training end condition is not reached, then based on the adjusted module parameters of the first preset neural network module, it returns to the first sample feature acquisition module for iterative training until the module parameters of the first preset neural network module adjusted by the first sample parameter adjustment module reach the preset training end condition and meet the determination condition of the first judgment module. Then, the module parameters of the first preset neural network module that meet the preset training end condition are input into the vehicle usage expectation prediction module for the area. This embodiment is based on the collected historical data, and through the first sample feature acquisition module, the vehicle usage expectation prediction module for the sample area, the first sample parameter adjustment module, and the first judgment module, a vehicle usage expectation prediction module for the area that meets the preset conditions is trained. This vehicle usage expectation prediction module for the area will be updated in real time during actual application. Through the method of online learning, it can accurately predict the distribution of passenger vehicle usage demand, promote the distribution of vehicle resources to be more in line with the user's demand expectation for vehicle resources, and achieve reasonable resource scheduling.

[0134] It should be noted that the training of the vehicle usage expectation prediction module for the area is an iterative training based on historical data. Only when the module parameters of the first preset neural network module reach the preset training end condition can the module operation parameters of the vehicle usage expectation prediction module for the area be obtained.

[0135] It should also be noted that both the first moment and the second moment obtained in the first sample feature acquisition module are subsequent moments based on the first moment and the second moment obtained when the first sample feature acquisition module was last run.

[0136] It should also be noted that the vehicle usage label information is obtained by weighted summation of the reserved vehicle usage information, the actual vehicle usage information, and the corresponding weights. Among them, the reserved vehicle usage information can specifically be the number of reserved vehicle usage orders, and the actual vehicle usage information can specifically be the number of actual completed vehicle usage orders. That is, the vehicle usage label information can be obtained by weighted summation of the number of reserved vehicle usage orders, the number of actual completed vehicle usage orders, and the corresponding weights.

[0137] Specifically, the embodiment of the present application may further include the training process of the dispatching area prediction module, and this process specifically includes:

[0138] The second sample feature acquisition module is used to acquire the sample vehicle features of each sample vehicle in the sample dispatching area at the third moment, and the dispatching label information corresponding to the sample dispatching area; the dispatching label information is determined according to the order receiving information of the dispatched vehicles in the sample dispatching area at the fourth moment, and the third moment refers to the moment after the fourth moment;

[0139] The sample vehicle usage expectation information extraction module is used to extract the vehicle usage expectation information corresponding to the sample dispatching area from the sample area vehicle usage expectation prediction module;

[0140] The sample feature fusion module is used for the fusion features corresponding to each sample vehicle, and the fusion features corresponding to each sample vehicle are based on the sample vehicle features of each sample vehicle and the vehicle usage expectation information corresponding to the sample dispatching area;

[0141] The sample dispatching area prediction module is used to obtain the dispatching area prediction results corresponding to each sample vehicle based on the fusion features corresponding to each sample vehicle and the second preset neural network module;

[0142] The second sample parameter adjustment module is used to adjust the module parameters of the second preset neural network module based on the difference between the dispatching area prediction results corresponding to each sample vehicle and the dispatching label information;

[0143] The second judgment module is used to compare the difference between the dispatching area prediction results corresponding to each sample vehicle and the dispatching label information, and adjust the module parameters of the second preset neural network module according to this difference.

[0144] In a specific embodiment, various sample vehicle characteristics at each historical moment within the sample scheduling area are collected in advance, as well as the scheduling label information corresponding to the sample scheduling area, that is, the order receiving information of the vehicles to be scheduled at each historical moment; the sample vehicle characteristics of each sample vehicle at the third moment within the sample scheduling area are obtained through the second sample feature acquisition module, as well as the scheduling label information corresponding to the sample scheduling area, that is, the order receiving information of the vehicles to be scheduled at the second moment; the vehicle usage expectation information corresponding to the sample scheduling area is extracted from the sample area vehicle usage expectation prediction module through the sample vehicle usage expectation information extraction module; the fusion features corresponding to each sample vehicle are obtained through the sample feature fusion module; the scheduling area prediction results corresponding to each sample vehicle are obtained through the sample scheduling area prediction module, that is, the predicted scheduling areas corresponding to each sample vehicle; the second sample parameter adjustment module compares the differences between the scheduling area prediction results corresponding to each sample vehicle and the scheduling label information, and adjusts the module parameters of the second preset neural network module according to the differences, where the second preset neural network module is a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm; the second judgment module determines whether the module parameters of the second preset neural network module adjusted by the second sample parameter adjustment module reach the preset training end condition. If the preset training end condition is not reached, based on the adjusted module parameters of the second preset neural network module, it returns to the second sample feature acquisition module for iterative training until the module parameters of the second preset neural network module adjusted by the second sample parameter adjustment module reach the preset training end condition and meet the determination condition of the second judgment module. Then, the module parameters of the second preset neural network module that meet the preset training end condition are input into the scheduling area prediction module. This embodiment is based on the collected historical data, and through the second sample feature acquisition module, the sample vehicle usage expectation information extraction module, the sample feature fusion module, the sample scheduling area prediction module, the second sample parameter adjustment module, and the second judgment module, a scheduling area prediction module that meets the preset conditions is trained. The scheduling area prediction module will be updated in real time during actual application. Through the online learning method, it can accurately predict the vehicle scheduling target area, promote the distribution of vehicle resources to better meet the user's demand expectations for vehicle resources, and achieve reasonable resource scheduling.

[0145] It should be noted that the training of the scheduling area prediction module is an iterative training based on historical data. Only when the module parameters of the second preset neural network module reach the preset training end condition can the module operation parameters of the scheduling area prediction module be obtained.

[0146] It should also be noted that during the iterative training process, both the third moment and the fourth moment obtained in the second sample feature acquisition module are subsequent moments based on the third moment and the fourth moment obtained in the previous run of the second sample feature acquisition module.

[0147] It should also be noted that the scheduling label information is obtained by the weighted sum of the order receiving volume, the vehicle usage settlement information, and the corresponding weights. Among them, the vehicle usage settlement information can be specifically the vehicle usage settlement amount, that is, the scheduling label information is obtained by the weighted sum of the order receiving volume, the vehicle usage settlement amount, and the corresponding weights.

[0148] Specifically, the first preset neural network module and the second preset neural network module adopt a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm.

[0149] In a specific embodiment, using a deep neural network or a multi-layer feedforward neural network based on the error backpropagation algorithm can represent complex functions with fewer parameters and reduce error signals, which is beneficial to ensuring the stability and accuracy of the operation of the first preset neural network model and the second preset neural network model, and further ensuring the stability and accuracy of the operation of the regional vehicle usage expectation prediction model and the scheduling area prediction model.

[0150] It should be noted that when the device provided in the above embodiment realizes its functions, only the above-mentioned division of each functional module is used for illustration. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the device provided in the above embodiment and the method embodiment belong to the same concept, and the specific implementation process can be seen in the method embodiment, which will not be elaborated here.

[0151] The vehicle scheduling device in the embodiment of the present application adopts the technical solution of a dual-intelligent module of a regional vehicle usage expectation prediction module and a scheduling area prediction module, and both the regional vehicle usage expectation prediction module and the scheduling area prediction module are deep reinforcement learning systems that are updated in real time, which is more suitable for simulating and predicting various uncertain factors in complex scenarios, making the prediction results closer to the real vehicle usage scenario, so as to achieve the effect of accurately predicting the distribution of passenger vehicle usage needs and the target area of vehicle scheduling, promoting the distribution of vehicle resources to better meet the user's demand expectations for vehicle resources, and realizing reasonable resource scheduling.

[0152] The embodiment of the present application also provides an electronic device, including a processor and a memory. At least one instruction or at least one program is stored in the memory, and at least one instruction or at least one program is loaded and executed by the processor to implement the vehicle scheduling method provided in the above method embodiment.

[0153] The memory can be used to store software programs and modules. By running the software programs and modules stored in the memory, the processor can execute various functional applications and vehicle scheduling. The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store the operating system, application programs required for functions, etc.; the data storage area can store data created according to the use of the device, etc. In addition, the memory can include high-speed random access memory and can also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage devices. Accordingly, the memory can also include a memory controller to provide the processor with access to the memory.

[0154] The method embodiments provided in the embodiments of the present application can be executed on a computer terminal, a server, or a similar computing device, that is, the above-mentioned electronic device can include a computer terminal, a server, or a similar computing device. Figure 5 It is a hardware structure block diagram of an electronic device that runs a vehicle scheduling method provided in the embodiments of the present application. As Figure 5 shown, the internal structure of the electronic device can include, but is not limited to: a processor, a network interface, and a memory. Among them, the processor, network interface, and memory in the electronic device can be connected through a bus or other means. In the embodiments of this specification Figure 5 it is taken as an example of being connected through a bus.

[0155] Among them, the processor (or CPU (Central Processing Unit, central processor)) is the computing core and control core of the electronic device. The network interface can optionally include a standard wired interface, a wireless interface (such as WI-FI, mobile communication interface, etc.). The memory (Memory) is the memory device in the electronic device, used to store programs and data. It can be understood that the memory here can be a high-speed RAM storage device or a non-unstable storage device (non-volatile memory), such as at least one magnetic disk storage device; optionally, it can also be at least one storage device located far from the aforementioned processor. The memory provides a storage space that stores the operating system of the electronic device, which can include, but is not limited to: Windows system (an operating system), Linux (an operating system), Android (a mobile operating system) system, IOS (a mobile operating system) system, etc. The present application does not make any limitations in this regard; and, one or more instructions suitable for being loaded and executed by the processor are also stored in this storage space. These instructions can be one or more computer programs (including program codes). In the embodiments of this specification, the processor loads and executes one or more instructions stored in the memory to implement the vehicle scheduling method provided in the above method embodiments.

[0156] An embodiment of the present application also provides a computer-readable storage medium, in which at least one instruction or at least one program segment is stored, and the at least one instruction or at least one program segment is loaded and executed by a processor to implement the vehicle scheduling method provided by the method embodiment.

[0157] Optionally, in this embodiment, the above storage medium may include but is not limited to: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), external hard drives, magnetic disks, or optical discs that can store program codes.

[0158] It should be noted that: the above sequence of embodiments of the present application is only for description and does not represent the superiority or inferiority of the embodiments. And the above specific embodiments of this specification have been described. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be executed in a different order than in the embodiments and still achieve the desired results. Additionally, the processes depicted in the drawings do not necessarily require the specific order or sequential order shown to achieve the desired results. In certain embodiments, multi-vehicle scheduling and parallel processing are also possible or may be advantageous.

[0159] Each embodiment in this specification is described in a progressive manner. The same or similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0160] Those of ordinary skill in the art can understand that all or part of the steps to implement the above embodiments can be completed by hardware, or can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium, and the above-mentioned storage medium can be a read-only memory, a magnetic disk, or an optical disc, etc.

[0161] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A vehicle scheduling method, characterized in that, Including: Obtain the operation status information of each operating vehicle within the scheduling area; Determine the operating vehicles in the idle state according to the operation status information to obtain the target vehicles within the scheduling area; Obtain the passenger characteristics within the scheduling area and the vehicle characteristics corresponding to each target vehicle; The scheduling area is any one of a plurality of scheduling areas obtained by dividing a preset geographical range; Input the passenger characteristics within the scheduling area into the regional vehicle usage expectation prediction model for vehicle usage expectation prediction processing to obtain the vehicle usage expectation information of the scheduling area; Fuse the vehicle characteristics corresponding to each target vehicle within the scheduling area with the vehicle usage expectation information corresponding to the scheduling area respectively to obtain the fusion characteristics corresponding to each target vehicle within the scheduling area; Input the fusion characteristics corresponding to each target vehicle within the scheduling area into the scheduling area prediction model for scheduling area prediction processing to obtain the predicted scheduling areas of each target vehicle within the scheduling area; Schedule the target vehicles in the plurality of scheduling areas according to the predicted scheduling areas of each target vehicle in each scheduling area; The training method of the regional vehicle usage expectation prediction model includes: Obtain the sample passenger characteristics of the sample scheduling area at the first moment, and the reservation vehicle usage information and actual vehicle usage information of the sample scheduling area at the second moment; the second moment refers to the moment after the first moment; Determine the weights corresponding to the reservation vehicle usage information and the actual vehicle usage information respectively; Perform weighted summation according to the reservation vehicle usage information, the actual vehicle usage information and the corresponding weights to obtain the vehicle usage label information; Input the sample passenger characteristics of the sample scheduling area into the first preset neural network model for vehicle usage expectation prediction to obtain the vehicle usage expectation information corresponding to the sample scheduling area; Adjust the model parameters of the first preset neural network model according to the difference between the vehicle usage expectation information corresponding to the sample scheduling area and the vehicle usage label information corresponding to the sample scheduling area; Continue iterative training based on the adjusted model parameters until the preset training end condition is reached to obtain the regional vehicle usage expectation prediction model; The training method of the scheduling area prediction model includes: Obtain the sample vehicle characteristics of each sample vehicle in the sample scheduling area at the third moment, and the scheduling label information corresponding to the sample scheduling area; the scheduling label information is determined according to the order receiving information of the scheduled vehicle in the sample scheduling area at the fourth moment, and the third moment refers to the moment after the fourth moment; Fuse the sample vehicle characteristics of each sample vehicle with the vehicle usage expectation information corresponding to the sample scheduling area to obtain the fusion characteristics corresponding to each sample vehicle, and input the fusion characteristics corresponding to each sample vehicle into the second preset neural network model for scheduling area prediction to obtain the scheduling area prediction results corresponding to each sample vehicle; Adjust the model parameters of the second preset neural network model according to the difference between the scheduling area prediction results corresponding to each sample vehicle and the scheduling label information; Continue iterative training based on the adjusted model parameters until the preset training end condition is reached, and obtain the scheduling area prediction model.

2. The vehicle scheduling method according to claim 1, wherein The method further includes: obtaining the order receiving information of the vehicles to be scheduled in the sample scheduling area at the fourth moment, where the order receiving information includes the order receiving volume and the vehicle use settlement information; Determine the weights corresponding to the order receiving volume and the vehicle use settlement information respectively; Perform weighted summation according to the order receiving volume, the vehicle use settlement information, and the corresponding weights to obtain the scheduling label information.

3. The vehicle scheduling method according to claim 1, wherein The first preset neural network model and the second preset neural network model are deep neural networks or multi-layer feedforward neural networks based on the error backpropagation algorithm.

4. A vehicle scheduling device, characterized in that, The device includes: A feature acquisition module, configured to acquire the operation status information of each operating vehicle in the scheduling area; determine the operating vehicles in the idle state according to the operation status information to obtain the target vehicles in the scheduling area; acquire the passenger characteristics in the scheduling area and the vehicle characteristics corresponding to each target vehicle, where the scheduling area is any one of multiple areas obtained by dividing a preset geographical range; A regional vehicle use expectation prediction module, configured to acquire the vehicle use expectation information of the scheduling area based on the passenger characteristics in the scheduling area; A vehicle use expectation information extraction module, configured to extract the vehicle use expectation information of the scheduling area; A feature fusion module, configured to acquire the fusion characteristics corresponding to each target vehicle in the scheduling area, where the fusion characteristics corresponding to each target vehicle in the scheduling area are based on the vehicle characteristics of each target vehicle in the scheduling area and the vehicle use expectation information corresponding to the scheduling area; A scheduling area prediction module, configured to acquire the predicted scheduling areas of each target vehicle in the scheduling area based on the fusion characteristics corresponding to each target vehicle in the scheduling area; A target vehicle scheduling module, configured to schedule the target vehicles in the multiple scheduling areas based on the predicted scheduling areas of each target vehicle in the scheduling area.

5. An electronic device, characterized in that, It includes a processor and a memory, and at least one instruction or at least one program is stored in the memory, and the at least one instruction or the at least one program is loaded and executed by the processor to implement the vehicle scheduling method according to any one of claims 1 to 3.

6. A computer-readable storage medium, in which at least one instruction or at least one program is stored, and the at least one instruction or the at least one program is loaded and executed by a processor to implement the vehicle scheduling method according to any one of claims 1 to 3.

Images