CN109146109B

CN109146109B - Order distribution and model training method and device

Info

Publication number: CN109146109B
Application number: CN201710458654.2A
Authority: CN
Inventors: 付俊强; 曾显越; 刘养彪
Original assignee: Beijing Didi Infinity Technology and Development Co Ltd
Current assignee: Beijing Didi Infinity Technology and Development Co Ltd
Priority date: 2017-06-16
Filing date: 2017-06-16
Publication date: 2021-07-27
Anticipated expiration: 2037-06-16
Also published as: CN109146109A

Abstract

The invention provides a method and a device for order distribution and model training, and relates to the technical field of machine learning, wherein the specific implementation mode of the method comprises the following steps: acquiring target information, wherein the target information comprises information of a service provider, information that the service provider has received a first ride-sharing order, information of a second ride-sharing order to be distributed and current real-time information; according to the target information, determining whether the second ride-sharing order is matched with the service provider or not by adopting a pre-trained target model; and if the second ride order matches the service provider, assigning the second ride order to the service provider. The embodiment enables the matching of the ride-sharing order and the service provider to be more reasonable, improves the service efficiency and also improves the utilization rate of service resources.

Description

Order distribution and model training method and device

Technical Field

The disclosure relates to the technical field of machine learning, in particular to a method and a device for order distribution and model training.

Background

In recent years, with the continuous development of internet technology, Offline business opportunities are combined with the internet, and a new O2O (Online To Offline) business model appears, so that the internet becomes an Offline transaction platform. Currently, O2O has entered a high-speed development stage, in which the O2O service of vehicles is the more successful O2O service developed. Taking the vehicle service as an example, currently, the vehicle service may include a plurality of different categories, such as a express service, a special car service, a tailgating service, a test driving service, and a rental car service. There are some types of vehicular services that can also provide pool services. When the ride-sharing service is provided, since a driver is required to provide services to two or more passengers at the same time, a plurality of starting points and a plurality of ending points of a journey are involved, and therefore, if the order matching is not reasonable, problems (such as too many detours or too low response rate) may be caused, so that not only the service efficiency is reduced, but also the utilization rate of service resources is reduced.

Disclosure of Invention

In order to solve the technical problem, the present disclosure provides a method and an apparatus for order allocation and model training.

According to a first aspect of the embodiments of the present disclosure, there is provided an order allocation method, including:

acquiring target information, wherein the target information comprises information of a service provider, information that the service provider has received a first ride-sharing order, information of a second ride-sharing order to be distributed and current real-time information;

according to the target information, determining whether the second ride-sharing order is matched with the service provider or not by adopting a pre-trained target model;

and if the second ride order matches the service provider, assigning the second ride order to the service provider.

According to a second aspect of the embodiments of the present disclosure, there is provided a training method of an order allocation model, including:

obtaining sample information, wherein the sample information comprises associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events;

training a target model by using the sample information;

the corresponding associated information comprises corresponding real-time information, information of a service provider in the historical ride-sharing event, information of a first historical ride-sharing order received by the service provider in the historical ride-sharing event first and information of a second historical ride-sharing order received later, aiming at any historical ride-sharing event.

According to a third aspect of the embodiments of the present disclosure, there is provided an order distribution apparatus including:

the system comprises an acquisition unit, a processing unit and a processing unit, wherein the acquisition unit is configured to acquire target information, and the target information comprises information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be distributed and current real-time information;

a determining unit configured to determine whether the second ride order matches the service provider by using a pre-trained goal model according to the goal information;

an assigning unit configured to assign the second ride order to the service provider when the second ride order matches the service provider.

According to a fourth aspect of the embodiments of the present disclosure, there is provided an order distribution apparatus including:

the acquisition unit is configured to acquire sample information, wherein the sample information comprises associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events;

a training unit configured to train out a target model using the sample information;

the service provider receives the information of a first historical ride-sharing order in the historical ride-sharing event firstly and receives the information of a second historical ride-sharing order later.

According to a fifth aspect of embodiments of the present disclosure, there is provided a computer storage medium having stored therein program instructions, the instructions comprising:

determining whether the second ride-sharing order is matched with the service provider or not by adopting a pre-trained target model according to the target information;

According to a sixth aspect of embodiments of the present disclosure, there is provided a computer storage medium having stored therein program instructions, the instructions comprising:

training a target model by using the sample information;

According to a seventh aspect of the embodiments of the present disclosure, there is provided an electronic apparatus including:

a processor adapted to implement instructions; and

a storage device adapted to store a plurality of instructions, the instructions adapted to be loaded and executed by a processor to:

According to an eighth aspect of embodiments of the present disclosure, there is provided an electronic apparatus including:

a processor adapted to implement instructions; and

training a target model by using the sample information;

The technical scheme provided by the embodiment of the disclosure can have the following beneficial effects:

according to the order distribution method and device provided by the embodiment of the disclosure, whether the second ride-sharing order is matched with the service provider or not is determined by acquiring the target information and adopting a pre-trained target model according to the target information. And when the second ride share order matches the service provider, assigning the second ride share order to the service provider. The target information comprises information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be distributed and current real-time information. Therefore, the matching between the ride-sharing order and the service provider is more reasonable, the service efficiency is improved, and the utilization rate of service resources is also improved.

According to the training method and device for the order distribution model provided by the embodiment of the disclosure, the target model is trained by obtaining the sample information and adopting the sample information, wherein the sample information comprises the associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events. Therefore, a model which can be used for allocating the ride-sharing orders is obtained, the match between the ride-sharing orders and the service provider is more reasonable, the service efficiency is improved, and the utilization rate of service resources is also improved.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and together with the description, serve to explain the principles of the disclosure.

FIG. 1 is a schematic diagram of an exemplary system architecture to which embodiments of the present disclosure may be applied;

FIG. 2 is a flow chart illustrating a method of allocating an order according to an exemplary embodiment of the present disclosure;

FIG. 3 is a flow chart illustrating another method of allocation of orders according to an exemplary embodiment of the present disclosure;

FIG. 4 is a flow chart illustrating a method of training an allocation model for an order according to an exemplary embodiment of the present disclosure;

FIG. 5 is a flow chart illustrating another method of training an allocation model for orders according to an exemplary embodiment of the present disclosure;

FIG. 6 is a block diagram of an order distribution apparatus according to an exemplary embodiment of the present disclosure;

FIG. 7 is a block diagram of a training apparatus for an order allocation model according to an exemplary embodiment of the present disclosure.

Detailed Description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the exemplary embodiments below are not intended to represent all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

The terminology used in the present disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used in this disclosure and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.

It is to be understood that although the terms first, second, third, etc. may be used herein to describe various information, such information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the present disclosure. The word "if" as used herein may be interpreted as "at … …" or "when … …" or "in response to a determination", depending on the context.

Referring to fig. 1, an exemplary system architecture diagram to which embodiments of the present disclosure may be applied:

as shown in fig. 1, system architecture 100 may include terminal devices, such as the illustrated

terminal devices

101, 102, 103, network 104, and server 105. It should be understood that the number or types of terminal devices, networks, and servers in fig. 1 are merely illustrative. There may be any number or type of terminal devices, networks, and servers, as desired for an implementation.

The network 104 is used to provide a medium for communication links between terminal devices and servers. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables, to name a few.

The

terminal devices

101, 102, 103 may interact with the server via the network 104 to receive or transmit requests or information or the like. The

terminal devices

101, 102, 103 may be various electronic devices including, but not limited to, smart phones, tablet computers, smart wearable devices, and personal digital assistants, among others.

The server 105 may be a server that provides various services. The server may store, analyze, and the like the received data, or may transmit a control command or a request to the terminal device or another server. The server may provide the service in response to a service request of the user. It will be appreciated that one server may provide one or more services, and that the same service may be provided by multiple servers.

The present disclosure will be described in detail with reference to specific examples.

As shown in fig. 2, fig. 2 is a flowchart illustrating a method for allocating an order, which may be applied to a server, according to an exemplary embodiment. The method comprises the following steps:

in step 201, destination information is obtained, where the destination information includes information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be allocated, and current real-time information.

In this embodiment, the service involved may be a car-to-car service (e.g., a car-to-car service, etc.), and the scenario involved may be a scenario in which a service provider providing the car composition service has received one car-to-car order and is waiting to receive another car-to-car order. For example, for a vehicle pool service, the service provider may be a driver carrying passengers. The first ride order is the first ride order received by the service provider, and the second ride order is the to-be-distributed ride order.

In this embodiment, the information of the service provider may include various information capable of characterizing the personal characteristics of the service provider. Taking the car ride-sharing service as an example, the service provider is a driver providing the car ride-sharing service, and the information of the service provider may include, but is not limited to, ID information of the driver, sex information of the driver, age information of the driver, service score information of the driver, star-level information of the driver, vehicle type information of the driver, current location information of the driver, and the like.

In this embodiment, the information of the first ride order may include various information included in the first ride order, for example, the information of the first ride order may include, but is not limited to, position information of a starting point and an ending point of a journey corresponding to the first ride order, an order-placing time corresponding to the first ride order, user information corresponding to the first ride order, and the like. The information of the second ride order may include various information included in the second ride order, for example, the information of the second ride order may include, but is not limited to, position information of a starting point and an ending point of a journey corresponding to the second ride order, an order-placing time corresponding to the second ride order, user information corresponding to the second ride order, and the like. The user information may include, but is not limited to, ID information of the user, user profile information of the user (e.g., sex information, age information, hobby information, professional information, etc.), and the like.

In the present embodiment, the current real-time information may include, but is not limited to, current weather information, current time information (e.g., time of day information, day of the week information, date of the gregorian calendar, holiday information, etc.), current traffic condition information, and the like.

In step 202, according to the objective information, it is determined whether the second ride share order matches the service provider using a pre-trained objective model.

In this embodiment, the target model is a pre-trained model, and the target model may be any one of the following: an XGBoost (Extreme Gradient Boosting) model; a linear regression model; a deep neural network DNN. It is to be understood that the target model may be other types of models, and the specific type of target model is not limited in this application.

In this embodiment, it may be determined whether the second ride together order matches the service provider by using a pre-trained goal model according to the goal information as follows: and acquiring target characteristic information based on the target information, inputting the target characteristic information into a target model, and acquiring matching parameters output by the target model. And if the matching parameter is larger than or equal to the preset threshold value, determining that the second ride together order is matched with the service provider.

In step 203, the second ride order is assigned to the service provider if the second ride order matches the service provider.

According to the order allocation method provided by the above embodiment of the present disclosure, by acquiring the target information, whether the second ride-sharing order is matched with the service provider is determined by using a pre-trained target model according to the target information. And when the second ride share order matches the service provider, assigning the second ride share order to the service provider. The target information comprises information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be distributed and current real-time information. Therefore, the matching between the ride-sharing order and the service provider is more reasonable, the service efficiency is improved, and the utilization rate of service resources is also improved.

FIG. 3 is a flow chart illustrating another method for allocating orders that details the process of determining whether a second ride share order matches a service provider using a pre-trained goal model based on goal information according to an exemplary embodiment that may be implemented in a server. The method may comprise the steps of:

in step 301, destination information is obtained, where the destination information includes information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be allocated, and current real-time information.

In step 302, target feature information is acquired based on the target information.

In this embodiment, the target feature information may include first target feature information and second target feature information, where the first target feature information is feature information that can be obtained directly based on the target information, and the second target feature information is feature information that needs to be estimated from the target information. Specifically, the first target feature information may be directly extracted from the target information, and estimation may be performed based on the target information (estimation may be performed by using a preset algorithm, strategy, or model) to obtain the second target feature information.

In this embodiment, the first target feature information may include one or more of the following: the service provider comprises gender characteristic information of the service provider, age characteristic information of the service provider, service score characteristic information of the service provider, star characteristic information of the service provider, vehicle type characteristic information of the service provider, current position characteristic information of the service provider, weather characteristic information and time characteristic information.

In this embodiment, the second target feature information may include one or more of the following: estimating distance characteristic information of an original journey corresponding to the first ride-sharing order; estimating distance characteristic information of an original journey corresponding to the second ride-sharing order; estimating distance characteristic information of a journey corresponding to the first ride-sharing order after ride sharing; estimating distance characteristic information of a journey corresponding to the second ride-sharing order after ride sharing; the first ride sharing order and the second ride sharing order correspond to the ride sharing estimated distance characteristic information after ride sharing; the first ride sharing order and the second ride sharing order correspond to ride sharing estimated time characteristic information after ride sharing; estimating detour distance characteristic information corresponding to the first ride-sharing order; estimating detour distance characteristic information corresponding to the second ride-sharing order; estimating detour time characteristic information corresponding to the first ride-sharing order; estimated detour time characteristic information corresponding to the second ride-sharing order; characteristic information of a ratio of an estimated detour distance corresponding to the first ride-sharing order to an original travel estimated distance corresponding to the first ride-sharing order; characteristic information of the ratio of the estimated detour distance corresponding to the second ride-sharing order and the original travel estimated distance corresponding to the second ride-sharing order; estimating characteristic information of the driving pickup time corresponding to the second ride-sharing order; estimating the characteristic information of the driving receiving distance corresponding to the second ride-sharing order; and the characteristic information of the ratio of the estimated driving receiving distance and the estimated travel distance corresponding to the second ride-sharing order.

In step 303, target feature information is input into the target model.

In step 304, matching parameters of the target model output are obtained.

In this embodiment, the target feature information may be input into the target model, and the matching parameter output by the target model may be obtained, where the matching parameter may represent a matching degree of the second ride share order and the service provider. And if the matching parameter is larger than or equal to the preset threshold value, the second ride-sharing order is matched with the service provider.

In step 305, if the matching parameter is greater than or equal to the preset threshold, it is determined that the second ride order matches the service provider.

In step 306, the second ride order is assigned to the service provider.

It should be noted that, for the same steps as in the embodiment of fig. 2, details are not repeated in the embodiment of fig. 3, and related contents may refer to the embodiment of fig. 2.

According to the order allocation method provided by the above embodiment of the present disclosure, the target information is acquired, the target characteristic information is acquired based on the target information, the target characteristic information is input into the target model, and the matching parameter output by the target model is acquired. And if the matching parameter is larger than or equal to the preset threshold value, determining that the second ride-sharing order is matched with the service provider, and distributing the second ride-sharing order to the service provider. Therefore, the matching between the ride-sharing order and the service provider is more reasonable, and the utilization rate of service resources is improved.

As shown in fig. 4, fig. 4 is a flowchart illustrating a method for training an order distribution model, which may be applied in a server, according to an exemplary embodiment. The method comprises the following steps:

in step 401, sample information is obtained, where the sample information includes associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events.

In this embodiment, for any historical ride-sharing event, the associated information corresponding to the historical ride-sharing event may include real-time information corresponding to the historical ride-sharing event, information of a service provider in the historical ride-sharing event, information of a first historical ride-sharing order received by the service provider in the historical ride-sharing event first, and information of a second historical ride-sharing order received later.

In this embodiment, the information of the service provider may include various information capable of characterizing the personal characteristics of the service provider. Taking the car ride-sharing service as an example, the service provider is a driver providing the car ride-sharing service, and the information of the service provider may include, but is not limited to, ID information of the driver, sex information of the driver, age information of the driver, service score information of the driver, star information of the driver, vehicle type information of the driver, position information of the driver, and the like.

In this embodiment, the information of the first historical ride order may include various information included in the first historical ride order, for example, the information of the first historical ride order may include, but is not limited to, position information of a starting point and an ending point of a journey corresponding to the first historical ride order, an order-issuing time corresponding to the first historical ride order, user information corresponding to the first historical ride order, and the like. The information of the second historical ride order may include various information included in the second historical ride order, for example, the information of the second historical ride order may include, but is not limited to, position information of a starting point and an ending point of a journey corresponding to the second historical ride order, an order-issuing time corresponding to the second historical ride order, user information corresponding to the second historical ride order, and the like. The user information may include, but is not limited to, ID information of the user, user profile information of the user (e.g., sex information, age information, hobby information, professional information, etc.), and the like.

In this embodiment, the real-time information corresponding to the historical ride-sharing event may include, but is not limited to, weather information, time information (e.g., time of day information, week information, date information of a gregorian calendar, date information of a lunar calendar, festival information, etc.), traffic condition information, and the like, when the historical ride-sharing event occurs.

In step 402, a target model is trained using the sample information.

In this embodiment, the target model may be any one of the following: an XGBoost (Extreme Gradient Boosting) model; a linear regression model; a deep neural network DNN. It is to be understood that the target model may be other types of models, and the specific type of target model is not limited in this application.

In this embodiment, first, a sample attribute corresponding to each historical ride-share event may be determined based on the sample information, where the sample attribute includes a positive sample attribute and a negative sample attribute, and then, sample feature information corresponding to each historical ride-share event may be obtained based on the sample information. And finally, training a target model according to the sample attribute and the sample characteristic information corresponding to each historical ride-sharing event.

According to the training method for the order distribution model provided by the above embodiment of the disclosure, the target model is trained by obtaining the sample information and adopting the sample information, wherein the sample information includes the associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events. Therefore, a model which can be used for allocating the ride-sharing orders is obtained, the match between the ride-sharing orders and the service provider is more reasonable, the service efficiency is improved, and the utilization rate of service resources is also improved.

Fig. 5 is a flowchart illustrating another method for training an order allocation model according to an exemplary embodiment, which describes in detail a process of training a target model using sample information, and may be applied to a server. The method may comprise the steps of:

in step 501, sample information is obtained, where the sample information includes associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events.

In step 502, sample attributes corresponding to each historical ride-share event are determined based on the sample information, the sample attributes including a positive sample attribute and a negative sample attribute.

In this embodiment, the sample attribute corresponding to each historical ride-sharing event may be determined according to the evaluation information and the response condition of the order in the sample information. For example, the sample attribute corresponding to the historical ride-sharing event that the order response is successful is a positive sample attribute. The sample attribute corresponding to the historical ride-sharing event with poor order evaluation and failed order response can be a negative sample attribute. The specific dividing mode of the positive and negative sample attributes is not limited in the application.

In step 503, sample feature information corresponding to each historical ride-share event is obtained based on the sample information.

In this embodiment, for any historical ride-sharing event, the sample feature information corresponding to the historical ride-sharing event may include first sample feature information and corresponding second sample feature information corresponding to the historical ride-sharing event. The first sample feature information is feature information that can be obtained directly based on the sample information, and the second sample feature information is feature information that needs to be estimated from the sample information. Specifically, the first sample feature information may be directly extracted from the sample information, and an estimation may be performed based on the sample information (the estimation may be performed by using a preset algorithm, strategy, or model) to obtain the second sample feature information.

In this embodiment, the first sample characteristic information may include one or more of the following: the service provider corresponding to the historical ride-sharing event comprises gender characteristic information of the service provider corresponding to the historical ride-sharing event, age characteristic information of the service provider corresponding to the historical ride-sharing event, service division characteristic information of the service provider corresponding to the historical ride-sharing event, star-level characteristic information of the service provider corresponding to the historical ride-sharing event, vehicle type characteristic information of the service provider corresponding to the historical ride-sharing event, location characteristic information of the service provider corresponding to the historical ride-sharing event, weather characteristic information corresponding to the historical ride-sharing event and time characteristic information corresponding to the historical ride-sharing event.

In this embodiment, the second sample feature information may include one or more of the following: estimating distance characteristic information of an original journey corresponding to the first historical ride-sharing order; estimating distance characteristic information of an original journey corresponding to the second historical ride-sharing order; estimating distance characteristic information of a journey corresponding to the first historical ride-sharing order after ride sharing; estimating distance characteristic information of a journey corresponding to the second historical ride order after ride combination; the first ride sharing order and the second ride sharing order correspond to the ride sharing estimated distance characteristic information after ride sharing; the first ride sharing order and the second ride sharing order correspond to ride sharing estimated time characteristic information after ride sharing; estimating detour distance characteristic information corresponding to the first historical ride-sharing order; estimated detour distance characteristic information corresponding to the second historical ride-sharing order; estimated detour time characteristic information corresponding to the first historical ride-sharing order; estimated detour time characteristic information corresponding to the second historical ride-sharing order; characteristic information of a ratio of an estimated detour distance corresponding to the first historical ride-sharing order to an original travel estimated distance corresponding to the first historical ride-sharing order; characteristic information of a ratio of an estimated detour distance corresponding to the second historical ride-sharing order to an original travel estimated distance corresponding to the second historical ride-sharing order; estimating characteristic information of the driving receiving moment corresponding to the second historical ride-sharing order; estimated driving receiving distance characteristic information corresponding to the second historical ride combination order; and the characteristic information of the ratio of the estimated driving receiving distance and the estimated travel distance corresponding to the second historical ride-sharing order.

In step 504, a target model is trained according to the sample attribute and the sample characteristic information corresponding to each historical ride-sharing event.

In this embodiment, the target model may be trained as follows: first, sample feature information of a data set is obtained, wherein each data set comprises a training set and a verification set (wherein the training set corresponds to a plurality of historical ride-sharing events, and the verification set corresponds to a plurality of historical ride-sharing events). And then, adjusting the parameters of the current model by adopting the sample characteristic information of the training set. And verifying the previously trained model by using the sample characteristic information of the verification set. And taking the current model as a trained model until the verification result meets the requirement.

The adjusting the parameters of the current model by using the sample feature information of the training set may include: and inputting the sample characteristic information of the training set into the current model to obtain a probability value corresponding to each historical ride-sharing event (namely, the probability that the sample attribute corresponding to the historical ride-sharing event is the positive sample attribute) as a reference parameter corresponding to the historical ride-sharing event. And obtaining an ROC curve according to the reference parameters corresponding to the plurality of historical ride-sharing events and the sample attributes corresponding to the historical ride-sharing events. And obtaining a corresponding AUC value based on the ROC curve. And if the AUC value is less than or equal to the preset threshold, adjusting the parameters of the current model, and then repeatedly executing the step of adjusting the parameters of the model. If the AUC value is greater than a preset threshold, a step of verifying the previously trained model is performed.

The verifying the previously trained model by using the sample feature information of the verification set may include: and inputting the sample characteristic information of the training set into the model trained in the front to obtain a corresponding first AUC value. And inputting the sample characteristic information of the verification set into the previously trained model to obtain a corresponding second AUC value. And subtracting the second AUC value from the first AUC value to obtain a difference, and repeating the step of adjusting the parameters of the model if the absolute value of the difference is greater than a preset threshold value. If the absolute value of the difference is smaller than the preset threshold, the verification result meets the requirement.

In the training method for the order allocation model provided by the above embodiment of the present disclosure, sample information is obtained, where the sample information includes associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events, and a sample attribute corresponding to each historical ride-sharing event is determined based on the sample information. And acquiring sample characteristic information corresponding to each historical ride-sharing event based on the sample information, and training a target model according to the sample attribute corresponding to each historical ride-sharing event and the sample characteristic information. Therefore, a model which can be used for allocating the ride-sharing orders is obtained, the matching between the ride-sharing orders and the service provider is further more reasonable, and the utilization rate of service resources is improved.

It should be noted that while the operations of the disclosed methods are depicted in the drawings in a particular order, this does not require or imply that these operations must be performed in this particular order, or that all of the illustrated operations must be performed, to achieve desirable results. Rather, the steps depicted in the flowcharts may change the order of execution. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.

Corresponding to the embodiment of the order distribution and model training method, the disclosure further provides an embodiment of an order distribution and model training device.

As shown in fig. 6, fig. 6 is a block diagram of an order distribution apparatus according to an exemplary embodiment of the present disclosure, the apparatus including: an acquisition unit 601, a determination unit 602, and an assignment unit 603.

The obtaining unit 601 is configured to obtain target information, where the target information includes information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be allocated, and current real-time information.

A determining unit 602, configured to determine whether the second ride order matches the service provider by using a pre-trained goal model according to the goal information.

An assigning unit 603 configured to assign the second ride share order to the service provider when the second ride share order matches with the service provider.

According to the order distribution device provided by the above embodiment of the present disclosure, by acquiring the target information, and according to the target information, determining whether the second ride-sharing order is matched with the service provider by using a pre-trained target model. And when the second ride share order matches the service provider, assigning the second ride share order to the service provider. The target information comprises information of a service provider, information that the service provider has received a first ride order, information of a second ride order to be distributed and current real-time information. Therefore, the matching between the ride-sharing order and the service provider is more reasonable, the service efficiency is improved, and the utilization rate of service resources is also improved.

In some optional embodiments, the determining unit 602 may include: a first acquisition subunit, an input subunit, a second acquisition subunit and a determination subunit (not shown in the figure).

Wherein the first acquiring subunit is configured to acquire the target feature information based on the target information.

An input subunit configured to input the target feature information into the target model.

And the second acquisition subunit is configured to acquire the matching parameters output by the target model.

And the determining subunit is configured to determine that the second ride order is matched with the service provider when the matching parameter is greater than or equal to a preset threshold value.

The order distribution device provided by the above embodiment of the present disclosure obtains the target characteristic information based on the target information by obtaining the target information, inputs the target characteristic information into the target model, and obtains the matching parameter output by the target model. And if the matching parameter is larger than or equal to the preset threshold value, determining that the second ride-sharing order is matched with the service provider, and distributing the second ride-sharing order to the service provider. Therefore, the matching between the ride-sharing order and the service provider is more reasonable, and the utilization rate of service resources is improved.

In other alternative embodiments, the target characteristic information may include first target characteristic information and second target characteristic information.

The first acquisition subunit is configured to: the first target characteristic information is directly extracted from the target information, and estimation is performed based on the target information to obtain second target characteristic information.

In further alternative embodiments, the information for the first ride order may include: the position information of the starting point and the end point of the travel corresponding to the first ride-sharing order and the order-issuing time corresponding to the first ride-sharing order.

The information of the second ride order may include: the position information of the starting point and the end point of the travel corresponding to the second ride-sharing order and the order-issuing time corresponding to the second ride-sharing order.

In further alternative embodiments, the second target characteristic information includes one or more of: estimating distance characteristic information of an original journey corresponding to the first ride-sharing order; estimating distance characteristic information of an original journey corresponding to the second ride-sharing order; estimating distance characteristic information of a journey corresponding to the first ride-sharing order after ride sharing; estimating distance characteristic information of a journey corresponding to the second ride-sharing order after ride sharing; the first ride sharing order and the second ride sharing order correspond to the ride sharing estimated distance characteristic information after ride sharing; the first ride sharing order and the second ride sharing order correspond to ride sharing estimated time characteristic information after ride sharing; estimating detour distance characteristic information corresponding to the first ride-sharing order; estimating detour distance characteristic information corresponding to the second ride-sharing order; estimating detour time characteristic information corresponding to the first ride-sharing order; estimated detour time characteristic information corresponding to the second ride-sharing order; characteristic information of a ratio of an estimated detour distance corresponding to the first ride-sharing order to an original travel estimated distance corresponding to the first ride-sharing order; characteristic information of the ratio of the estimated detour distance corresponding to the second ride-sharing order and the original travel estimated distance corresponding to the second ride-sharing order; estimating characteristic information of the driving pickup time corresponding to the second ride-sharing order; estimating the characteristic information of the driving receiving distance corresponding to the second ride-sharing order; and the characteristic information of the ratio of the estimated driving receiving distance and the estimated travel distance corresponding to the second ride-sharing order.

In further alternative embodiments, the object model may include any one of: an extreme gradient ascending XGboost model; a linear regression model; a deep neural network DNN.

It should be understood that the above-mentioned device may be preset in the server, and may also be loaded into the server by downloading or the like. The corresponding units in the above-described apparatus may cooperate with units in the server to implement the allocation scheme for the order.

As shown in fig. 7, fig. 7 is a block diagram of a training apparatus for an order allocation model according to an exemplary embodiment of the present disclosure, the apparatus including: an acquisition unit 701 and a training unit 702.

The obtaining unit 701 is configured to obtain sample information, where the sample information includes associated information corresponding to each historical ride-sharing event in a plurality of historical ride-sharing events.

A training unit 702 configured to train out a target model using the sample information.

The training apparatus for the order distribution model provided in the above embodiment of the present disclosure trains the target model by obtaining sample information and using the sample information, where the sample information includes associated information corresponding to each historical ride share event in a plurality of historical ride share events. Therefore, a model which can be used for allocating the ride-sharing orders is obtained, the match between the ride-sharing orders and the service provider is more reasonable, the service efficiency is improved, and the utilization rate of service resources is also improved.

In further alternative embodiments, training unit 702 may include: a determining subunit, an acquiring subunit and a training subunit (not shown in the figure).

The determining subunit is configured to determine, based on the sample information, a sample attribute corresponding to each historical ride-sharing event, where the sample attribute includes a positive sample attribute and a negative sample attribute.

And the acquisition subunit is configured to acquire sample characteristic information corresponding to each historical ride-sharing event based on the sample information.

And the training subunit is configured to train a target model according to the sample attribute and the sample characteristic information corresponding to each historical ride-sharing event.

The verifying the previously trained model by using the sample feature information of the verification set may include: and inputting the sample characteristic information of the training set into the model trained in the front to obtain a corresponding first AUC value. And inputting the sample characteristic information of the verification set into the previously trained model to obtain a corresponding second AUC value. And subtracting the second AUC value from the first AUC value to obtain a difference, and if the absolute value of the difference is larger than a preset threshold value, readjusting the parameters of the model. If the absolute value of the difference is smaller than the preset threshold, the verification result meets the requirement.

The training device for the order distribution model provided by the above embodiment of the present disclosure obtains sample information, where the sample information includes associated information corresponding to each historical ride share event in a plurality of historical ride share events, and determines a sample attribute corresponding to each historical ride share event based on the sample information. And acquiring sample characteristic information corresponding to each historical ride-sharing event based on the sample information, and training a target model according to the sample attribute corresponding to each historical ride-sharing event and the sample characteristic information. Therefore, a model which can be used for allocating the ride-sharing orders is obtained, the matching between the ride-sharing orders and the service provider is further more reasonable, and the utilization rate of service resources is improved.

In some further alternative embodiments, for any historical ride-share event, the corresponding sample feature information includes first sample feature information corresponding to the historical ride-share event and second sample feature information corresponding to the historical ride-share event.

The obtaining subunit obtains the sample feature information corresponding to the historical ride-sharing event based on the sample information in the following manner: and directly extracting corresponding first sample characteristic information from the associated information corresponding to the historical ride-sharing event in the sample information, and estimating based on the associated information corresponding to the historical ride-sharing event in the sample information to obtain corresponding second target characteristic information.

In further alternative embodiments, the information for the first historical ride orders may include: the travel starting point and the travel end point corresponding to the first historical ride-sharing order, and the order-issuing time corresponding to the first historical ride-sharing order.

The information of the second historical ride order may include: the position information of the starting point and the end point of the journey corresponding to the second historical ride-sharing order and the order-issuing time corresponding to the second historical ride-sharing order.

In further alternative embodiments, for any historical ride-share event, the corresponding second sample characteristic information includes one or more of: estimating distance characteristic information of an original journey corresponding to the first historical ride-sharing order; estimating distance characteristic information of an original journey corresponding to the second historical ride-sharing order; estimating distance characteristic information of a journey corresponding to the first historical ride-sharing order after ride sharing; estimating distance characteristic information of a journey corresponding to the second historical ride order after ride combination; the first historical ride-sharing order and the second historical ride-sharing order correspond to estimated ride-sharing distance characteristic information after ride sharing; the first historical ride sharing order and the second historical ride sharing order correspond to ride sharing estimated time characteristic information after ride sharing; estimating detour distance characteristic information corresponding to the first historical ride-sharing order; estimated detour distance characteristic information corresponding to the second historical ride-sharing order; estimated detour time characteristic information corresponding to the first historical ride-sharing order; estimated detour time characteristic information corresponding to the second historical ride-sharing order; characteristic information of a ratio of an estimated detour distance corresponding to the first historical ride-sharing order to an original travel estimated distance corresponding to the first historical ride-sharing order; characteristic information of a ratio of an estimated detour distance corresponding to the second historical ride-sharing order to an original travel estimated distance corresponding to the second historical ride-sharing order; estimating characteristic information of the driving receiving moment corresponding to the second historical ride-sharing order; estimated driving receiving distance characteristic information corresponding to the second historical ride combination order; and the characteristic information of the ratio of the estimated driving receiving distance and the estimated travel distance corresponding to the second historical ride-sharing order.

In further alternative embodiments, the object model may include any one of: XGboost model; a linear regression model; a deep neural network DNN.

For the device embodiments, since they substantially correspond to the method embodiments, reference may be made to the partial description of the method embodiments for relevant points. The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the disclosed solution. One of ordinary skill in the art can understand and implement it without inventive effort.

It should be understood that the above-mentioned device may be preset in the server, and may also be loaded into the server by downloading or the like. The corresponding units in the above-described apparatus may cooperate with units in the server to implement a training scheme for the allocation model of the order.

Embodiments of the present disclosure may take the form of a computer program product embodied on one or more storage media including, but not limited to, disk storage, CD-ROM, optical storage, and the like, having program code embodied therein.

Accordingly, an embodiment of the present disclosure further provides a computer storage medium, in which program instructions are stored, where the instructions include:

Accordingly, another computer storage medium is provided in an embodiment of the present disclosure, in which program instructions are stored, and the instructions include:

training a target model by using the sample information;

The unit modules described in the embodiments of the present disclosure may be implemented by software or hardware. The described unit modules may also be provided in a processor, and may be described as: a processor includes an acquisition unit, a determination unit, and an allocation unit. Where the names of these unit modules do not constitute a limitation on the unit modules themselves under certain circumstances, for example, the determination unit may also be described as "a unit for determining whether the second ride order matches the service provider using a pre-trained goal model based on the goal information".

As another aspect, the present disclosure also provides a computer-readable storage medium, which may be the computer-readable storage medium included in the apparatus in the above-described embodiments; or it may be a computer-readable storage medium that exists separately and is not assembled into a terminal or server. The computer readable storage medium stores one or more programs for use by one or more processors in performing the order distribution, model training methods described in the present disclosure.

Computer-usable storage media include permanent and non-permanent, removable and non-removable media, and information storage may be implemented by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of the storage medium of the computer include, but are not limited to: phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technologies, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic tape storage or other magnetic storage devices, or any other non-transmission medium, may be used to store information that may be accessed by a computing device.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This disclosure is intended to cover any variations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It will be understood that the present disclosure is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for allocating an order, the method comprising:

assigning the second ride order to the service provider if the second ride order matches the service provider;

determining whether the second ride together order is matched with the service provider by adopting a pre-trained target model according to the target information, wherein the determining comprises the following steps:

acquiring target characteristic information based on the target information;

the target characteristic information comprises first target characteristic information and second target characteristic information;

the obtaining of the target feature information based on the target information includes:

directly extracting first target characteristic information from the target information; and

and estimating based on the target information to obtain the second target characteristic information.

2. The method of claim 1, wherein determining whether the second ride order matches the service provider using a pre-trained goal model based on the goal information further comprises:

inputting the target feature information into the target model;

acquiring matching parameters output by the target model;

and if the matching parameter is greater than or equal to a preset threshold value, determining that the second ride together order is matched with the service provider.

3. The method of claim 1, wherein the information of the first ride order comprises: position information of a travel starting point and a travel end point corresponding to the first ride-sharing order; the order issuing moment corresponding to the first ride sharing order;

the information of the second ride order comprises: position information of a travel starting point and a travel end point corresponding to the second ride-sharing order; and the order issuing moment corresponding to the second ride-sharing order.

4. The method of claim 3, wherein the second target feature information comprises one or more of:

estimating distance characteristic information of an original journey corresponding to the first ride-sharing order;

estimating distance characteristic information of an original journey corresponding to the second ride-sharing order;

estimating distance characteristic information of a journey corresponding to the first ride-sharing order after ride sharing;

estimating distance characteristic information of a journey corresponding to the second ride-sharing order after ride sharing;

the first ride sharing order and the second ride sharing order are corresponding to estimated ride sharing distance characteristic information after ride sharing;

the first ride sharing order and the second ride sharing order correspond to ride sharing estimated time characteristic information after ride sharing;

the estimated detour distance characteristic information corresponding to the first ride-sharing order;

the estimated detour distance characteristic information corresponding to the second ride-sharing order;

estimated detour time characteristic information corresponding to the first ride-sharing order;

estimated detour time characteristic information corresponding to the second ride-sharing order;

characteristic information of a ratio of an estimated detour distance corresponding to the first ride-sharing order to an original travel estimated distance corresponding to the first ride-sharing order;

characteristic information of a ratio of an estimated detour distance corresponding to the second ride-sharing order to an original travel estimated distance corresponding to the second ride-sharing order;

estimating characteristic information of the driving receiving time corresponding to the second ride sharing order;

estimated driving receiving distance characteristic information corresponding to the second ride-sharing order;

and the second ride-sharing order corresponds to characteristic information of the ratio of the estimated driving receiving distance to the estimated travel distance.

5. The method according to any of claims 1-4, wherein the object model comprises any of:

an extreme gradient ascending XGboost model;

a linear regression model;

a deep neural network DNN.

6. A method for training an order allocation model, the method comprising:

training a target model by using the sample information;

aiming at any historical ride-sharing event, the corresponding associated information comprises corresponding real-time information, information of a service provider in the historical ride-sharing event, information of a first historical ride-sharing order received by the service provider in the historical ride-sharing event firstly and information of a second historical ride-sharing order received later;

training a target model by using the sample information, including: acquiring sample characteristic information corresponding to each historical ride-sharing event based on the sample information;

for any historical ride-sharing event, the corresponding sample characteristic information comprises first sample characteristic information and second sample characteristic information corresponding to the historical ride-sharing event;

acquiring sample characteristic information corresponding to the historical ride-sharing event based on the sample information in the following way:

directly extracting corresponding first sample characteristic information from the associated information corresponding to the historical ride-sharing event in the sample information; and

and estimating based on the associated information corresponding to the historical ride-sharing event in the sample information to obtain the corresponding second target characteristic information.

7. The method of claim 6, wherein training out a target model using the sample information further comprises:

determining sample attributes corresponding to each historical ride-share event based on the sample information, wherein the sample attributes comprise positive sample attributes and negative sample attributes;

and training a target model according to the sample attribute and the sample characteristic information corresponding to each historical ride-sharing event.

8. The method of claim 6, wherein the information of the first historical ride order comprises: the position information of a travel starting point and a travel end point corresponding to the first historical ride-sharing order; the order issuing moment corresponding to the first historical ride-sharing order;

the information of the second historical ride order comprises: the position information of a travel starting point and a travel end point corresponding to the second historical ride-sharing order; and the order issuing time corresponding to the second historical ride-sharing order.

9. The method of claim 8, wherein for any historical ride-share event, the corresponding second sample characteristic information comprises one or more of:

estimating distance characteristic information of an original journey corresponding to the first historical ride-sharing order;

estimating distance characteristic information of an original journey corresponding to the second historical ride-sharing order;

estimating travel distance characteristic information corresponding to the first historical ride order after ride combination;

estimating travel distance characteristic information corresponding to the second historical ride order after ride combination;

the first historical ride sharing order and the second historical ride sharing order are subjected to ride sharing and estimation distance characteristic information corresponding to the ride sharing;

the first historical ride sharing order and the second historical ride sharing order correspond to ride sharing estimated time characteristic information after ride sharing;

estimated detour distance characteristic information corresponding to the first historical ride-sharing order;

estimated detour distance characteristic information corresponding to the second historical ride-sharing order;

estimated detour time characteristic information corresponding to the first historical ride-sharing order;

estimated detour time characteristic information corresponding to the second historical ride-sharing order;

characteristic information of a ratio of an estimated detour distance corresponding to the first historical ride-sharing order to an original travel estimated distance corresponding to the first historical ride-sharing order;

the characteristic information of the ratio of the estimated detour distance corresponding to the second historical ride-sharing order to the original travel estimated distance corresponding to the second historical ride-sharing order;

the estimated driving receiving time characteristic information corresponding to the second historical ride-sharing order;

estimated driving receiving distance characteristic information corresponding to the second historical ride-sharing order;

and the second historical ride-sharing order corresponds to characteristic information of the ratio of the estimated driving receiving distance to the estimated travel distance.

10. The method according to any of claims 6-9, wherein the object model comprises any of:

XGboost model;

a linear regression model;

a deep neural network DNN.

11. An apparatus for distributing an order, the apparatus comprising:

an allocation unit configured to allocate the second ride share order to the service provider when the second ride share order matches the service provider;

a first acquisition subunit configured to acquire target feature information based on the target information;

the first acquisition subunit is configured to:

12. The apparatus of claim 11, wherein the determining unit further comprises:

an input subunit configured to input the target feature information into the target model;

a second obtaining subunit configured to obtain a matching parameter output by the target model;

a determining subunit configured to determine that the second ride order matches the service provider when the matching parameter is greater than or equal to a preset threshold.

13. The apparatus of claim 11, wherein the information of the first ride order comprises: position information of a travel starting point and a travel end point corresponding to the first ride-sharing order; the order issuing moment corresponding to the first ride sharing order;

14. The apparatus of claim 13, wherein the second target feature information comprises one or more of:

15. The apparatus according to any of claims 11-14, wherein the object model comprises any of:

an extreme gradient ascending XGboost model;

a linear regression model;

a deep neural network DNN.

16. An apparatus for training an allocation model of an order, the apparatus comprising:

an obtaining subunit, configured to obtain, based on the sample information, sample feature information corresponding to each historical ride-sharing event;

the obtaining subunit obtains sample characteristic information corresponding to the historical ride-sharing event based on the sample information in the following manner:

17. The apparatus of claim 16, wherein the training unit further comprises:

a determining subunit configured to determine, based on the sample information, a sample attribute corresponding to each historical ride-sharing event, where the sample attribute includes a positive sample attribute and a negative sample attribute;

18. The apparatus of claim 16, wherein the information of the first historical ride order comprises: the position information of a travel starting point and a travel end point corresponding to the first historical ride-sharing order; the order issuing moment corresponding to the first historical ride-sharing order;

19. The apparatus of claim 18, wherein for any historical ride-share event, the corresponding second sample characteristic information comprises one or more of:

20. The apparatus according to any of claims 16-19, wherein the object model comprises any of:

XGboost model;

a linear regression model;

a deep neural network DNN.

21. A computer storage medium having program instructions stored therein, the instructions comprising:

acquiring target characteristic information based on the target information;

22. A computer storage medium having program instructions stored therein, the instructions comprising:

training a target model by using the sample information;

23. An electronic device, comprising:

a processor adapted to implement instructions; and

acquiring target characteristic information based on the target information;

24. An electronic device, comprising:

a processor adapted to implement instructions; and

training a target model by using the sample information;