Goods matching method and device, and electronic device

Abstract

The application provides a cargo-matching method and device and electronic equipment, and relates to the technical field of cargo transportation. The cargo-matching method comprises the following steps: determining cargo-matching information corresponding to a to-be-matched cargo source, wherein the cargo-matching information comprises the to-be-matched cargo source, a to-be-selected driver corresponding to the to-be-matched cargo source, a first cargo-matching score, a driver cargo source pool, and a second cargo-matching score corresponding to the driver cargo source pool; determining a cargo source selection probability of the to-be-selected driver selecting each cargo source based on the to-be-matched cargo source, the first cargo-matching score, the driver cargo source pool, the second cargo-matching score, and historical transaction orders of the to-be-selected driver; and determining a target driver matched with the to-be-matched cargo source from the to-be-selected driver based on the cargo-matching information and the cargo source selection probability, so that the sum of transaction probabilities of the to-be-matched cargo source and all cargo sources in the driver cargo source pool is maximum. The scheme of the application can reduce invalid exposure of the cargo source and save exposure resources of the cargo source.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of goods transportation, and in particular to a driver-goods matching method and device and electronic equipment. BACKGROUND

[0002] With the development of Internet technology, a digital freight platform emerges as the times require, which can integrate logistics supply and demand resources, provide information sharing of goods, consignors, freight drivers and service providers, and plays an important role in the logistics field. Freight drivers can use the digital freight platform to undertake goods transportation, wherein the matching between the driver and the goods source is an important link, and efficient and reasonable matching between the driver and the goods source is crucial to improve transportation efficiency and reduce costs.

[0003] In the related technology of driver-goods matching, when a driver comes to the freight platform to find goods, the freight platform can match the most suitable goods source to the driver according to a recommendation algorithm, and display these matched goods sources in the form of a list in the interface. However, this matching mechanism is based on the scoring of the goods source, and the higher the score of the high-quality goods source, the more likely it is to be displayed to more drivers, and the goods source has uniqueness, and only one driver can finally transact. This will make the high-quality goods source be exposed too much, causing waste of goods exposure resources. SUMMARY

[0004] The present application provides a driver-goods matching method, device and electronic equipment to solve the problem of waste of goods exposure resources in the related art driver-goods matching method.

[0005] In a first aspect, the present application provides a driver-goods matching method, comprising:

[0006] determining driver-goods information corresponding to a to-be-matched goods source, wherein the driver-goods information comprises the to-be-matched goods source, a to-be-selected driver corresponding to the to-be-matched goods source, a first driver-goods score, a driver-goods pool, and a second driver-goods score corresponding to the driver-goods pool;

[0007] determining a goods selection probability of the to-be-selected driver selecting each goods source based on the to-be-matched goods source, the first driver-goods score, the driver-goods pool, the second driver-goods score, and historical transaction orders of the to-be-selected driver;

[0008] determining a target driver matched with the to-be-matched goods source from the to-be-selected driver based on the driver-goods information and the goods selection probability, with the sum of transaction probabilities of the to-be-matched goods source and all goods sources in the driver-goods pool being maximum as the target.

[0009] Optionally, the target driver matched with the to-be-matched freight source is determined from the to-be-selected drivers based on the freight information and the freight source selection probability, with a sum of transaction probabilities of the to-be-matched freight source and all freight sources in the driver freight source pool being maximum as an objective function.

[0010] The objective function is constructed based on the freight information and the freight source selection probability, with a sum of transaction probabilities of the to-be-matched freight source and all freight sources in the driver freight source pool being maximum as an objective function.

[0011] The to-be-matched freight source is matched with all drivers in the to-be-selected drivers as an initial solution of the objective function, and a driver bringing a maximum gain value to the objective function in each round of traversal is subtracted by using a traversal pruning principle until no gain of the objective function is produced, and the finally reserved driver is determined as the target driver matched with the to-be-matched freight source.

[0012] Optionally, the to-be-matched freight source is matched with all drivers in the to-be-selected drivers as an initial solution of the objective function, and a driver bringing a maximum gain value to the objective function in each round of traversal is subtracted by using a traversal pruning principle until no gain of the objective function is produced, and the finally reserved driver is determined as the target driver matched with the to-be-matched freight source, comprising:

[0013] The to-be-matched freight source is matched with all drivers in the to-be-selected drivers as an initial solution of the objective function, and a driver bringing a maximum gain value to the objective function in each round of traversal is subtracted by using a traversal pruning principle until no gain of the objective function is produced, and the finally reserved driver is determined as the target driver matched with the to-be-matched freight source.

[0014] Each solution in the initial solution is traversed, and a gain value brought to the objective function value when the each solution is pruned is determined.

[0015] The maximum gain value in the gain value is determined, and the solution corresponding to the maximum gain value is pruned from the initial solution to obtain an updated initial solution.

[0016] The updated initial solution is taken as the initial solution of the objective function again until the gain values produced are all less than or equal to 0, and it is determined that no gain of the objective function is produced, and the driver corresponding to the finally reserved solution in the initial solution is determined as the target driver matched with the to-be-matched freight source.

[0017] Optionally, the objective function is represented as:

[0018] ;

[0019] wherein, P i represents a transaction probability of the i-th freight source, a function value representing the target function, D represents a set of the to-be-selected drivers, C represents a set of the to-be-matched goods sources, a goods source selection probability representing that the jth driver selects the ith goods source, a representation quantity representing whether the ith goods source is matched to the jth driver; when the ith goods source is not matched to the jth driver, and when the ith goods source is matched to the jth driver.

[0020] Optionally, the determining of the driver-goods information corresponding to the to-be-matched goods source comprises:

[0021] obtaining goods-to-car recall sorting information of the to-be-matched goods source, and drivers in the goods-to-car recall sorting information are sorted in descending order of matching degrees;

[0022] determining first preset numbers of drivers in the goods-to-car recall sorting information as the to-be-selected drivers corresponding to the to-be-matched goods source, and determining matching degrees corresponding to the to-be-selected drivers in the goods-to-car recall sorting information as first driver-goods scores of the to-be-selected drivers;

[0023] obtaining a goods finding list of the to-be-selected drivers, and goods sources in the goods finding list are sorted in descending order of driver-goods matching degrees;

[0024] determining first preset numbers of goods sources in the goods finding list as the driver-goods source pool, and determining driver-goods matching degrees corresponding to the goods sources in the goods finding list as second driver-goods scores corresponding to the driver-goods source pool;

[0025] determining the to-be-matched goods source, the to-be-selected drivers corresponding to the to-be-matched goods source and the first driver-goods scores, the driver-goods source pool and the second driver-goods scores corresponding to the driver-goods source pool as the driver-goods information corresponding to the to-be-matched goods source.

[0026] Optionally, the driver-goods matching method further comprises:

[0027] updating the driver-goods source pool and the second driver-goods scores corresponding to the driver-goods source pool in a case where a goods finding operation of the to-be-selected driver is detected;

[0028] in a case where a first goods source is offline and the first goods source is a goods source cached in the driver-goods source pool, deleting the first goods source from the driver-goods source pool.

[0029] Optionally, the determining of the goods source selection probability of the to-be-selected driver selecting each goods source based on the to-be-matched goods source, the first driver-goods score, the driver-goods source pool, the second driver-goods score and historical transaction orders of the to-be-selected driver comprises:

[0030] constructing the to-be-matched cargo source and the driver cargo source pool as a cargo source list;

[0031] taking each cargo source in the cargo source list as a target cargo source, obtaining a first utility value of the target cargo source being selected and a second utility value of the cargo source list not being selected; the first utility value and the second utility value are used to represent the attractiveness of the target cargo source;

[0032] determining a total expected utility value of all cargo sources in the cargo source list based on the first utility value and a fitting parameter;

[0033] based on the fitting parameter, the total expected utility value and the second utility value, determining a first probability of the to-be-selected driver selecting a cargo source in the cargo source list for a transaction and a second probability of the to-be-selected driver not selecting a cargo source in the cargo source list for a transaction;

[0034] for each target cargo source, determining a third probability of the target cargo source being selected based on the first utility value of the target cargo source;

[0035] based on the first probability, the second probability and the third probability, determining the cargo source selection probability of the to-be-selected driver selecting each cargo source;

[0036] wherein, in a case that the target cargo source is the to-be-matched cargo source, the first utility value corresponding to the target cargo source is the first driver-cargo score; in a case that the target cargo source is a cargo source in the driver cargo source pool, the first utility value corresponding to the target cargo source is the second driver-cargo score; and the fitting parameter is determined by maximum log-likelihood estimation on historical driver behavior data.

[0037] Optionally, the determining the cargo source selection probability of the to-be-selected driver selecting each cargo source based on the first probability, the second probability and the third probability comprises:

[0038] determining the cargo source selection probability of the to-be-selected driver selecting each cargo source based on the first probability, the second probability and the third probability by using the following formula:

[0039] ;

[0040] wherein, denotes the cargo source selection probability of the jth to-be-selected driver selecting each cargo source, denotes the first probability, denotes the second probability, denotes the third probability, denotes the cargo source list, representing a selection operation of the driver to be selected, representing a selection of no transaction.

[0041] In a second aspect, the present application provides a driver-goods matching device, comprising:

[0042] a first determining module configured to determine driver-goods information corresponding to a to-be-matched goods source, the driver-goods information comprising the to-be-matched goods source, a to-be-selected driver corresponding to the to-be-matched goods source, a first driver-goods score, a driver-goods pool, and a second driver-goods score corresponding to the driver-goods pool;

[0043] a second determining module configured to determine a goods source selection probability of the to-be-selected driver selecting each goods source based on the to-be-matched goods source, the first driver-goods score, the driver-goods pool, the second driver-goods score, and historical transaction orders of the to-be-selected driver;

[0044] a third determining module configured to determine a target driver matched with the to-be-matched goods source from the to-be-selected driver based on the driver-goods information and the goods source selection probability, with a sum of transaction probabilities of the to-be-matched goods source and all goods sources in the driver-goods pool as a target.

[0045] In a third aspect, the present application provides an electronic device, comprising a memory and a processor, the memory storing a computer program capable of running on the processor, and the processor implements steps of the driver-goods matching method according to any one of the first aspect when executing the computer program.

[0046] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program implements steps of the driver-goods matching method according to any one of the first aspect when executed by a processor.

[0047] The method, device and electronic equipment provided in the application determine the driver-goods information corresponding to the to-be-matched goods source first, the driver-goods information including the to-be-matched goods source, the to-be-selected driver corresponding to the to-be-matched goods source, the first driver-goods score, the driver-goods pool and the second driver-goods score corresponding to the driver-goods pool, then determine the goods source selection probability of the to-be-selected driver selecting each goods source based on the to-be-matched goods source, the first driver-goods score, the driver-goods pool, the second driver-goods score and the historical transaction order of the to-be-selected driver, and then determine the target driver matched with the to-be-matched goods source from the to-be-selected driver based on the driver-goods information and the goods source selection probability, so that the target driver matched with the to-be-matched goods source can be determined from the to-be-selected driver with the maximum transaction amount of all goods sources as the target, when the driver looks for goods in the goods transport platform, the to-be-matched goods source can be exposed only in the goods source recommendation list of the matched target driver, and other drivers will not see the to-be-matched goods source, thereby reducing invalid exposure and saving exposure resources of the goods source. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 A flowchart of a driver-goods matching method provided by an embodiment of the application is shown.

[0049] Figure 2 A structural diagram of a driver-goods matching device provided by an embodiment of the application is shown. DETAILED DESCRIPTION

[0050] In the application, "at least one" means one or more, and "multiple" means two or more. The "and / or" describes the association relationship of the associated objects, which means that there can be three kinds of relationships, for example, A and / or B, which can represent the following three cases: A exists alone, A and B exist together, and B exists alone. A and B can be singular or plural. The character " / " generally represents an "or" relationship between the front and rear associated objects. "At least one of the following" or similar expressions means any combination of these items, including single item or any combination of multiple items. For example, at least one of a, b or c alone can represent: a alone, b alone, c alone, combination of a and b, combination of a and c, combination of b and c, or combination of a, b and c. A, b and c can be single or multiple. In addition, the terms "first", "second" are only for description purposes and cannot be understood as indicating or implying relative importance.

[0051] The matching between the driver and the goods source (i.e., driver-goods matching) is an important part of the goods transport platform, and efficient and reasonable matching between the driver and the goods source is crucial for improving transportation efficiency and reducing costs.

[0052] In the related art, when a driver comes to a freight platform to find goods, the freight platform can match the most suitable goods source to the driver according to a recommendation algorithm, and display these matched goods in the form of a list in the interface. However, this matching mechanism is based on the scoring of the goods source, and the higher the quality of the goods source, the higher the score, and the more likely it is to be displayed to more drivers. However, the goods source is unique, and only one driver can eventually complete the transaction. This will cause the high-quality goods source to be exposed too much, not only wasting the exposure resources of the goods source, but also causing a clear contrast with the remaining ordinary goods source in the driver's goods list, causing mutual influence between the goods sources, making it difficult for other ordinary goods sources to be accepted by the driver, and thus affecting the transportation of goods. Therefore, it is necessary to further optimize the strategy of driver-goods matching from a more comprehensive perspective.

[0053] In the related art, a time window of orders can be cached, and then a vehicle-goods matching optimal decision is made with the current idle driver as the target of maximizing transaction volume, and then the order is dispatched according to the decision result. However, this method is only suitable for scenarios where orders are relatively dense and dispatching does not need to consider the contrast between goods sources, and requires the platform to have high control over the driver. However, for the scenario of goods transportation, due to the sparseness of the time and space distribution of the goods source and the high timeliness requirement of the goods source being accepted, it is impossible to achieve concentrated allocation of goods, and the sparseness of the time and space distribution of the goods source also determines that the driver mainly performs autonomous order acceptance when finding goods in the list, and it is difficult to dispatch orders. Therefore, this vehicle-goods matching strategy is not suitable for the driver-goods matching scenario of goods transportation.

[0054] Based on this, the embodiments of the present application provide a new driver-goods matching method, which can determine a target driver matched for a ticket of goods from the to-be-selected drivers when the ticket of goods is issued, with the goal of maximizing the transaction volume of all goods. When the driver comes to the freight platform to find goods, the to-be-matched goods source can be exposed only in the goods source recommendation list of the matched target driver, and other drivers will not be able to see the to-be-matched goods source, thereby reducing invalid exposure, saving the exposure resources of the goods source, and avoiding unnecessary contrast between the goods sources, reducing the mutual influence between the goods sources, and ensuring that all types of goods can be normally transported.

[0055] The driver-goods matching method provided by the embodiments of the present application can be applied to an electronic device, and can also be applied to a driver-goods matching device provided in the electronic device. The driver-goods matching device can be implemented by software, hardware, or a combination of both. The electronic device can include at least one of a server, a mobile phone, a computer, a vehicle terminal, a tablet computer, a wearable device, and the like, but is not limited thereto. The server can include at least one of a standalone server, a virtual server, and a cluster server, but is not limited thereto.

[0056] In the following, the driver-goods matching method provided by the embodiments of the present application is described in detail. Figure 1 The driver-goods matching method provided by the embodiments of the present application is described in detail.

[0057] Figure 1 A flowchart of a method for matching a cargo source with a driver is shown, and the method can include the following steps 110-130. Figure 1

[0058] Step 110: Determine the driver-cargo information corresponding to the cargo source to be matched, which includes the cargo source to be matched, the driver to be selected corresponding to the cargo source to be matched, the first driver-cargo score, the driver-cargo pool, and the second driver-cargo score corresponding to the driver-cargo pool.

[0059] The cargo source to be matched is a new cargo to be issued. The driver to be selected is the driver who most wants the new cargo. The driver to be selected can be determined according to the cargo-finder-call-back ordering information of the cargo source to be matched, for example, the top N (N is a positive integer) drivers with higher matching degrees are determined as the driver to be selected. The driver-cargo pool is used to save other cargos that the driver to be selected most wants, except for the cargo source to be matched. It can be understood that these cargos are the cargos that are most likely to be displayed simultaneously with the new cargo in the cargo list, and can be determined according to the cargo-finder list of all cargo-finder drivers in a preset time period, for example, the top M (M is a positive integer) cargos with higher frequencies are cached in the driver-cargo pool.

[0060] The first driver-cargo score is used to represent the matching degree of the cargo source to be matched and the driver to be selected, and the second driver-cargo score is used to represent the matching degree of the driver to be selected and each cargo in the driver-cargo pool.

[0061] Specifically, step 110 of determining the driver-cargo information corresponding to the cargo source to be matched can be implemented through the following steps 111-115.

[0062] Step 111: Obtain the cargo-finder-call-back ordering information of the cargo source to be matched. The drivers in the cargo-finder-call-back ordering information are sorted in descending order of matching degree.

[0063] When a cargo source to be matched (i.e., a new cargo) is published on a cargo platform, the cargo-finder-call-back information corresponding to the cargo source to be matched can be determined according to a preset cargo-finder-call-back condition. The cargo-finder-call-back information can include driver information and the matching degree of the driver and the cargo source to be matched. Then, the cargo-finder-call-back information can be sorted in descending order of matching degree to obtain the cargo-finder-call-back ordering information.

[0064] ​For example, the preset conditions for calling back the driver for the cargo can include the driver information of the drivers within a preset range of the departure location of the cargo to be matched, the distance between the location of the drivers and the departure location of the cargo to be matched, the destination information of the cargo to be matched, the city information of the cities where the drivers often drive, the vehicle type and length required by the cargo to be matched, the vehicle type and length of the drivers, the volume and weight of the cargo to be matched, and the vehicle load information of the drivers, and the like. Based on the information, the electronic device can analyze the matching degree of the cargo to be matched in different dimensions of the driver-cargo matching, for example, the closer the distance between the location of the driver and the departure location of the cargo to be matched, the higher the matching degree, the more similar the vehicle type and length of the driver and the vehicle type and length required by the cargo to be matched, the higher the matching degree, and the more suitable the volume and weight of the cargo to be matched and the vehicle load information of the driver, the higher the matching degree. The electronic device can comprehensively consider the matching degree in different dimensions, integrate the matching degree in different dimensions by using an algorithm such as weighted summation, and determine the matching degree of the driver and the cargo to be matched.

[0065] Step 112: determining the first preset number of drivers in the cargo-to-vehicle calling back sorting information as the to-be-selected drivers corresponding to the cargo to be matched, and determining the matching degrees of the to-be-selected drivers in the cargo-to-vehicle calling back sorting information as the first driver-cargo scores of the to-be-selected drivers.

[0066] The first preset number can be set according to actual application, for example, can be 5, 8, 10, and the like, which is not specially limited in the present application.

[0067] In the cargo-to-vehicle calling back sorting information, the first N (N is a positive integer) drivers can be determined as the to-be-selected drivers, which are the drivers with higher matching degrees with the cargo to be matched, and can be confirmed as the drivers who most want to transport the cargo to be matched. For each to-be-selected driver, the matching degree thereof can be determined as the corresponding first driver-cargo score, and the higher the first driver-cargo score, the greater the probability that the corresponding driver undertakes the cargo to be matched.

[0068] Step 113: obtaining the cargo list of the to-be-selected driver, and the cargos in the cargo list are sorted in descending order of the driver-cargo matching degree.

[0069] For each driver who comes to the cargo transportation platform to find cargos, the cargo transportation platform can display the corresponding cargo list to the driver, that is, the cargo transportation platform can maintain a cargo list for each driver, which can be updated regularly or irregularly according to the operation of the driver, the update of the cargo, and the like. Based on this, for each to-be-selected driver who comes to the cargo transportation platform to find cargos, the cargo list of the to-be-selected driver can be directly obtained.

[0070] Step 114: determining the first second preset number of the goods sources in the goods finding list as the driver goods source pool, and determining the driver-goods matching degrees of the goods sources in the goods finding list as the second driver-goods scores corresponding to the driver goods source pool.

[0071] The second preset number can be set according to actual application, for example, can be 5, 10, 15, etc., and the application does not make special limitation thereto.

[0072] For each to-be-selected driver, a driver goods source pool can be maintained for the to-be-selected driver, and the first second preset number of goods sources, for example, the first 10 goods sources, in the goods finding list of the to-be-selected driver are cached in the driver goods source pool. The goods sources in the driver goods source pool are the most matched goods sources when the to-be-selected driver finds goods, and the to-be-selected driver can confirm the goods sources as the most wanted goods sources of the to-be-selected driver except the to-be-matched goods sources.

[0073] In an embodiment, the driver-goods matching method can further include: in the case that the finding operation of the to-be-selected driver is detected, updating the driver goods source pool and the second driver-goods scores corresponding to the driver goods source pool.

[0074] Specifically, for each to-be-selected driver, when the to-be-selected driver finds goods on the freight platform each time, the electronic device can detect the finding operation of the to-be-selected driver through the freight platform, for example, detects the operation of calling out the goods finding page. At this time, the electronic device can obtain the first second preset number of goods sources in the goods finding list of the to-be-selected driver within a preset time period before the current time, and update the goods sources in the driver goods source pool of the to-be-selected driver to the first second preset number of goods sources obtained at this time, so as to refresh the cached goods sources and the corresponding second driver-goods scores in the driver goods source pool.

[0075] In this way, the goods sources are cached through the driver goods source pool, and the timeliness of the goods sources can be met without hiding the goods.

[0076] In an embodiment, the driver-goods matching method can further include: in the case that the first goods source is offline and the first goods source is a cached goods source in the driver goods source pool, deleting the first goods source from the driver goods source pool.

[0077] Specifically, the goods sources published on the freight platform are in a dynamic changing state, and can be offline at a certain moment, for example, have been accepted or withdrawn by the consignor. For a goods source, when the goods source is offline, if the goods source is a cached goods source in the driver goods source pool, the goods source is deleted from the driver goods source pool at the same time. In this way, the timeliness of the cached goods sources in the driver goods source pool can be ensured.

[0078] Step 115: determining the to-be-matched cargo source, the to-be-selected driver corresponding to the to-be-matched cargo source and the first driver-cargo score, the driver-cargo pool, and the second driver-cargo score corresponding to the driver-cargo pool as the driver-cargo information corresponding to the to-be-matched cargo source.

[0079] After determining the to-be-selected driver corresponding to the to-be-matched cargo source and the first driver-cargo score, the driver-cargo pool, and the second driver-cargo score corresponding to the driver-cargo pool, the to-be-matched cargo source, the to-be-selected driver corresponding to the to-be-matched cargo source and the first driver-cargo score, the driver-cargo pool, and the second driver-cargo score corresponding to the driver-cargo pool can be determined as the driver-cargo information corresponding to the to-be-matched cargo source, for example, these information can be constituted into a driver-cargo matrix and represented in the form of a matrix. The driver-cargo information can represent all drivers who most want the to-be-matched cargo source and the remaining cargo sources that these drivers want.

[0080] Step 120: determining the cargo source selection probability of the to-be-selected driver selecting each cargo source based on the to-be-matched cargo source, the first driver-cargo score, the driver-cargo pool, the second driver-cargo score, and the historical transaction order of the to-be-selected driver.

[0081] After determining the driver-cargo information corresponding to the to-be-matched cargo source, the to-be-matched cargo source, the first driver-cargo score, the driver-cargo pool, and the second driver-cargo score in the driver-cargo information, and the historical transaction order of the to-be-selected driver, the principle of the Nested Multinomial Logit (NMNL) model can be used to determine the cargo source selection probability of the to-be-selected driver selecting each cargo source.

[0082] As to whether the to-be-selected driver transacts the cargo source, it can be regarded as a two-stage decision, the to-be-selected driver can first decide whether to select the cargo source in the cargo source list, and then decide which cargo source to select.

[0083] Based on this, for any one selection of the jth to-be-selected driver , the cargo source selection probability of the to-be-selected driver can be expressed as the following formula (1):

[0084] (1)

[0085] Wherein, represents the cargo source selection probability of the jth to-be-selected driver selecting each cargo source, represents the cargo source list, o represents the selection operation of the to-be-selected driver, represents not selecting to transact, represents the first probability of the to-be-selected driver selecting the cargo source in the cargo source list to transact, represents the second probability of the to-be-selected driver not selecting the cargo source in the cargo source list to transact, represents the third probability of the to-be-selected driver selecting the cargo source selected by the selection operation o in the cargo source list.

[0086] According to the principle shown in formula (1), in one embodiment, step 120 determines the probability of the driver to select each cargo source based on the cargo source to be matched, the first driver cargo score, the driver cargo source pool, the second driver cargo score and the historical transaction orders of the driver to be selected. This can be achieved through the following steps 121 to 126.

[0087] Step 121: Construct a cargo list from the cargo sources to be matched and the driver cargo source pool.

[0088] The drivers to be selected are the drivers who most want to be matched with freight. For each driver to be selected, the corresponding driver freight pool is the most suitable freight for that driver when looking for freight, excluding the freight to be matched. The freight to be matched and the driver freight pool can be used to construct the freight list of the drivers to be selected.

[0089] Step 122: Take each source in the source list as the target source, and obtain the first utility value of the target source when it is selected and the second utility value of the source list when it is not selected.

[0090] The first utility value and the second utility value are used to characterize the attractiveness or value of the target goods. The higher the first utility value, the higher the attractiveness of the target goods and the greater the probability of being selected; the higher the second utility value, the lower the attractiveness of each target goods in the goods list and the lower the probability of being selected.

[0091] Specifically, the cargo list includes cargo to be matched and cargo in the driver's cargo pool. When the target cargo is cargo to be matched, the first utility value corresponding to the target cargo is the first driver-cargo score; when the target cargo is cargo in the driver's cargo pool, the first utility value corresponding to the target cargo is the second driver-cargo score.

[0092] The second utility value for the driver not selecting cargo from the cargo source list can be determined based on the driver's historical completed orders. For example, it can be the average utility value of those historical completed orders. Specifically, the second utility value can be determined according to the following formula (2):

[0093] (2)

[0094] in, This represents the second utility value, where m represents the number of historical orders. This represents the driver-cargo matching degree between the k-th historical completed order of the j-th candidate driver and the j-th candidate driver. This driver-cargo matching degree is also the utility value.

[0095] Step 123: Determine the total expected utility value of all goods in the goods list based on each first utility value and the fitted parameters.

[0096] The fitting parameter is determined by maximum log-likelihood estimation on historical driver behavior data. The historical driver behavior data can include information of drivers who have transacted with the transacted freight sources in the historical freight list, information of drivers who have not transacted with the untransacted freight sources, and matching degrees between the drivers and the freight sources.

[0097] Specifically, the fitting parameter can be determined by the log of the likelihood function shown in the following formula (3):

[0098] (3)

[0099] wherein, the likelihood function is represented by, the fitting parameter is represented by; whether the driver and the freight source transact or not can be determined based on the driver information and the corresponding transacted freight sources and untransacted freight sources in the historical driver behavior data, and when it represents transacting, and when it represents not transacting; D represents a set of drivers to be selected; the freight source selection probability is represented by, the freight list is represented by, the selection operation of the driver to be selected is represented by, and the selection of not transacting is represented by.

[0100] After the fitting parameter and the first utility value of each freight source in the freight list are obtained, the total expected utility value of all freight sources in the freight list can be determined according to the following formula (4):

[0101] (4)

[0102] wherein, the total expected utility value is represented by, the freight list is represented by, i represents the i-th freight source in the freight list, and the first utility value is represented by.

[0103] Step 124: Based on the fitting parameter, the total expected utility value, and the second utility value, a first probability that the driver to be selected transacts with the freight source in the freight list and a second probability that the driver to be selected does not transact with the freight source in the freight list are determined.

[0104] After the fitting parameter , the total expected utility value , and the second utility value are determined, the fitting parameter , the total expected utility value , and the second utility value The first probability of the to-be-selected driver selecting a cargo in the cargo source list for transaction is determined according to formula (5) as follows:

[0105] (5)

[0106] wherein, The first probability of the to-be-selected driver selecting a cargo in the cargo source list for transaction is represented.

[0107] The second probability of the to-be-selected driver not selecting a cargo in the cargo source list for transaction can be determined according to formula (6) as follows:

[0108] (6)

[0109] wherein, The second probability of the to-be-selected driver not selecting a cargo in the cargo source list for transaction is represented.

[0110] Step 125: For each target cargo source, a third probability of the target cargo source being selected is determined based on the first utility value of the target cargo source.

[0111] Specifically, for each target cargo source, the third probability of the target cargo source being selected can be determined according to formula (7) as follows based on the first utility value of the target cargo source:

[0112] (7)

[0113] wherein, The third probability of the target cargo source being selected is represented, The first utility value of the i th target cargo source in the cargo source list is represented.

[0114] Step 126: Based on the first probability, the second probability and the third probability, a cargo source selection probability of the to-be-selected driver selecting each cargo source is determined.

[0115] Specifically, after the first probability, the second probability and the third probability are determined, the cargo source selection probability of the to-be-selected driver selecting each cargo source can be determined according to formula (1) as above.

[0116] In this way, by using the NMNL model principle to determine the cargo source selection probability of the to-be-selected driver selecting each cargo source based on the to-be-matched cargo source in the driver-cargo information, the first driver-cargo score, the driver-cargo pool and the second driver-cargo score and the historical transaction order of the to-be-selected driver, the selection behavior of the to-be-selected driver can be quantified, the mutual influence between the cargo sources is considered, and the coupling problem of "displaying the cargo source set to affect the selection probability" is avoided.

[0117] Step 130: based on the cargo information and the cargo selection probability, a target driver matched with the to-be-matched cargo is determined from the to-be-selected drivers, with the sum of the transaction probabilities of all cargos in the driver cargo pool being maximized as the target.

[0118] After determining the cargo selection probability of each cargo in the to-be-selected driver selection cargo list based on the cargo information and the cargo selection probability, a target function can be established based on the cargo information and the cargo selection probability, with the sum of the transaction probabilities of all cargos in the driver cargo pool being maximized as the target, and the target driver matched with the to-be-matched cargo is determined from the to-be-selected drivers by optimizing and solving the target function. In this way, when the driver comes to the cargo platform to find the cargo, the to-be-matched cargo can be exposed only in the cargo recommendation list of the target driver matched, and other drivers will not be able to see the to-be-matched cargo, thereby reducing invalid exposure and saving exposure resources of the cargo.

[0119] Specifically, step 130 determines the target driver matched with the to-be-matched cargo from the to-be-selected drivers based on the cargo information and the cargo selection probability, with the sum of the transaction probabilities of all cargos in the driver cargo pool being maximized as the target, which can be achieved through steps 131-132 as follows.

[0120] Step 131: based on the cargo information and the cargo selection probability, a target function is constructed with the sum of the transaction probabilities of all cargos in the driver cargo pool being maximized as the target.

[0121] Specifically, all cargos in the to-be-matched cargo and the driver cargo pool can form a cargo list. After obtaining the cargo information and the cargo selection probability, for the i-th cargo in the cargo list, the transaction probability of the i-th cargo can be determined based on the cargo selection probability of the to-be-selected driver selecting the cargo according to the following formula (8):

[0122] (8)

[0123] wherein P i represents the transaction probability of the i-th cargo, D represents the set of to-be-selected drivers, represents the cargo selection probability of the j-th driver selecting the i-th cargo, represents the representation of whether the i-th cargo is matched to the j-th driver; when , it means that the i-th cargo is not matched to the j-th driver, and when , it means that the i-th cargo is matched to the j-th driver.

[0124] Further, the sum of the transaction probabilities of all cargos in the cargo list can be obtained, and then a target function is constructed with the sum of the transaction probabilities being maximized as the target. The constructed target function can be represented by the following formula (9):

[0125] (9)

[0126] wherein, represents a function value of the objective function, C represents a set of driver source pools and to-be-matched source, P i represents a transaction probability of the i-th source, D represents a set of to-be-selected drivers, represents a source selection probability of the j-th driver selecting the i-th source, represents a representation quantity of whether the i-th source is matched to the j-th driver; when represents that the i-th source is not matched to the j-th driver, when represents that the i-th source is matched to the j-th driver.

[0127] Step 132: taking matching the to-be-matched source to all drivers in the to-be-selected driver as an initial solution of the objective function, subtracting a driver bringing the maximum gain value to the objective function in each round of traversal by using the traversal pruning principle until no gain of the objective function is produced, and determining the finally reserved driver as a target driver matched to the to-be-matched source. The traversal pruning principle refers to that in the process of traversing the initial solution in each round, whether the initial solution is pruned from the initial solution set after the end of the round is determined by judging the influence of the initial solution on the objective function after the initial solution is subtracted, so as to optimize and solve the objective function.

[0128] wherein, each solution in the initial solution represents a driver-source pair, and is used to represent a matching pair of the to-be-matched source and the to-be-selected driver.

[0129] Specifically, step 132 takes matching the to-be-matched source to all drivers in the to-be-selected driver as an initial solution of the objective function, subtracts a driver bringing the maximum gain value to the objective function in each round of traversal by using the traversal pruning algorithm until no gain of the objective function is produced, and determines the finally reserved driver as a target driver matched to the to-be-matched source, which can be realized through steps 1321~step 1325 as follows.

[0130] Step 1321: taking matching the to-be-matched source to all drivers in the to-be-selected driver as an initial solution of the objective function, and determining a target function value of the objective function based on the initial solution.

[0131] For the to-be-matched source c1, the to-be-matched source c1 can be matched to each driver in the set D of to-be-selected drivers, so as to construct an initial solution X of the objective function, which can be represented as formula (10) as follows:

[0132] (10)

[0133] wherein, represents a characteristic quantity indicating whether the to-be-matched cargo source c1 is matched to the jth driver in the set D of to-be-selected drivers, represents that the to-be-matched cargo source c1 is matched to the jth driver in the set D of to-be-selected drivers. Then, for the initial solution X, it represents that the to-be-matched cargo source c1 is matched to each driver in the set D of to-be-selected drivers.

[0134] After the initial solution X is constructed, the objective function value of the objective function as shown in formula (9) can be determined according to the initial solution.

[0135] Step 1322: Each solution in the initial solution is traversed to determine the gain value brought to the objective function value when each solution is pruned.

[0136] Each solution in the initial solution represents a driver-cargo pair, and each solution in the initial solution can be regarded as an edge. All edges in the initial solution X are traversed to determine the gain value brought to the objective function value when each edge is pruned .

[0137] For example, assuming that the initial solution X includes 3 solutions, i.e., 3 edges, the first edge is pruned to determine the objective function value, and the objective function value is compared with the objective function value obtained in step 1321 to obtain the first gain value of the objective function value after the first edge is pruned. Similarly, the second gain value of the objective function value after the second edge is pruned and the third gain value of the objective function value after the third edge is pruned can be obtained.

[0138] Step 1323: The maximum gain value in the gain values is determined, and the solution corresponding to the maximum gain value is pruned from the initial solution to obtain an updated initial solution.

[0139] For example, assuming that the initial solution X includes 3 solutions in step 1322, the first gain value, the second gain value, and the third gain value are obtained, and the maximum gain value is determined therefrom, for example, the first gain value. If the first gain value is greater than 0, it indicates that a gain is generated, and at this time, the first edge (i.e., the first solution) corresponding to the first gain value can be pruned from the initial solution to obtain an updated initial solution. At this time, the updated initial solution X includes the remaining 2 solutions.

[0140] Step 1324: The updated initial solution is re-used as the initial solution of the objective function.

[0141] Step 1325: The above steps 1321 to 1324 are repeatedly executed until the generated gain value is less than or equal to 0, and it is determined that no gain of the objective function is generated. The driver corresponding to the solution finally retained in the initial solution is determined as the target driver to which the to-be-matched cargo source is matched.

[0142] For example, as illustrated in step 1323, the initial solution X includes the remaining two solutions, which can be used again as the initial solutions for the objective function. Then, steps 1321 to 1324 are repeated until the gain value after subtracting each edge is less than or equal to 0, indicating that no further gain is generated for the objective function. This signifies that the sum of the transaction probabilities of all freight sources in the driver-freight pool has reached its maximum. At this point, the final solution retained in the initial solution is the optimal solution, and the driver corresponding to each solution is the target driver matched with the freight to be matched. When a driver comes to the freight platform to find freight, the freight to be matched can only be exposed in the freight recommendation list of these matched target drivers. Other drivers will not see the freight to be matched, effectively reducing invalid exposure of the freight to be matched and saving freight exposure resources.

[0143] Based on steps 1321 to 1325 above, the principle can be described as follows:

[0144] Traversal For all edges, determine the gain value that removing each edge brings to the objective function. Then, the maximum gain value Δmax is determined from these gain values, and the edge corresponding to the maximum gain value Δmax is obtained. At this point, check if Δmax is greater than 0. If it is, it indicates that a gain has occurred in this round of traversal, and then cut the edge from X. X updated to ,in, This means subtracting the number of drivers from the set D of drivers to be selected. Next, continue iterating through the remaining edges in X to determine the gain value that removing each edge from X brings to the objective function. Then, determine the maximum gain value Δmax from these gain values ​​and obtain the edge corresponding to that maximum gain value Δmax. If Δmax is greater than 0, remove that edge from X to update X. Continue updating X in this loop until Δmax is less than or equal to 0, indicating that it no longer contributes to the objective function's gain. At this point, the sum of the transaction probabilities of all cargo in the driver-cargo pool reaches its maximum, and the drivers retained in X can be identified as the target drivers matching the cargo to be matched.

[0145] The method for matching a driver and a cargo provided in the embodiments of the present application can determine the driver-cargo information corresponding to the cargo to be matched, the driver-cargo information including the cargo to be matched, a driver to be selected corresponding to the cargo to be matched and a first driver-cargo score, a driver-cargo pool and a second driver-cargo score corresponding to the driver-cargo pool, then determine the cargo selection probability of the driver to be selected selecting each cargo based on the cargo to be matched, the first driver-cargo score, the driver-cargo pool, the second driver-cargo score and historical transaction orders of the driver to be selected, and then determine the target driver matched with the cargo to be matched from the driver to be selected based on the driver-cargo information and the cargo selection probability, with the sum of transaction probabilities of the cargo to be matched and all cargos in the driver-cargo pool being maximum as the target. In this way, the target driver matched with the cargo to be matched can be determined from the driver to be selected with the transaction volume of all cargos being maximized as the target, when the driver comes to the cargo transportation platform to find cargos, the cargo to be matched can be exposed only in the cargo recommendation list of the target driver matched, and other drivers will not see the cargo to be matched, thereby reducing invalid exposure and saving exposure resources of the cargos. Moreover, unnecessary comparison between cargos is avoided, mutual influence between cargos is reduced, and all types of cargos can be normally transported.

[0146] The method for matching a driver and a cargo provided in the embodiments of the present application can construct the driver-cargo information based on the real-time cached driver-cargo pool, without covering cargos, and can meet the timeliness requirement of the cargos. In addition, the driver and cargo matching decision is made with the transaction volume of all cargos being maximized as the target, thereby effectively reducing invalid exposure, improving the overall transaction volume and improving the efficiency of the cargo transportation platform.

[0147] The embodiments of the present application further provide a device for matching a driver and a cargo, Figure 2 The structure schematic diagram of the device for matching a driver and a cargo provided in the embodiments of the present application is shown, and the device for matching a driver and a cargo can include: Figure 2

[0148] The first determining module 210 is configured to determine the driver-cargo information corresponding to the cargo to be matched, the driver-cargo information including the cargo to be matched, a driver to be selected corresponding to the cargo to be matched and a first driver-cargo score, a driver-cargo pool and a second driver-cargo score corresponding to the driver-cargo pool;

[0149] The second determining module 220 is configured to determine the cargo selection probability of the driver to be selected selecting each cargo based on the cargo to be matched, the first driver-cargo score, the driver-cargo pool, the second driver-cargo score and historical transaction orders of the driver to be selected;

[0150] The third determining module 230 is configured to determine the target driver matched with the cargo to be matched from the driver to be selected based on the driver-cargo information and the cargo selection probability, with the sum of transaction probabilities of the cargo to be matched and all cargos in the driver-cargo pool being maximum as the target.

[0151] In one embodiment, the third determining module 230 can include:​

[0152] a target function construction unit, configured to construct a target function based on the cargo information and the cargo source selection probability, so as to maximize the sum of the transaction probabilities of all cargo sources in the cargo source pool to which the cargo source is to be matched and the drivers;

[0153] a target driver determination unit, configured to use the traversal pruning principle to subtract a driver that brings the maximum gain value to the target function in each round of traversal from an initial solution of the target function that matches the cargo source to be matched to all drivers in the driver selection, until no gain of the target function is generated, and determine the finally retained driver as the target driver to which the cargo source to be matched is matched.

[0154] In an embodiment, the target driver determination unit is specifically configured to: determine a target function value of the target function based on an initial solution of the target function that matches the cargo source to be matched to all drivers in the driver selection; traverse each solution in the initial solution, and determine a gain value brought to the target function value when each solution is pruned; determine a maximum gain value in the gain values, and prune the solution corresponding to the maximum gain value from the initial solution to obtain an updated initial solution; use the updated initial solution as the initial solution of the target function again until the generated gain values are all less than or equal to 0, determine that no gain of the target function is generated, and determine a driver corresponding to a finally retained solution in the initial solution as the target driver to which the cargo source to be matched is matched.

[0155] In an embodiment, the constructed target function can be represented as:

[0156] ;

[0157] wherein P i represents a transaction probability of the ith cargo source, represents a function value of the target function, D represents a set of the driver selection, and C represents a set of the driver cargo pool and the cargo source to be matched, represents a cargo source selection probability of the jth driver selecting the ith cargo source, represents a representation quantity of whether the ith cargo source is matched to the jth driver; when , it represents that the ith cargo source is not matched to the jth driver, and when , it represents that the ith cargo source is matched to the jth driver.

[0158] In an embodiment, the first determination module 210 can include:

[0159] a first acquisition unit, configured to acquire cargo-finding-vehicle recall sorting information of the cargo source to be matched, and the drivers in the cargo-finding-vehicle recall sorting information are sorted in descending order of matching degree;

[0160] The first determining unit is configured to determine the first preset number of drivers in the car-hunting-car recall ranking information as to-be-selected drivers corresponding to to-be-matched goods sources, and determine the matching degrees of the to-be-selected drivers in the car-hunting-car recall ranking information as first driver-goods scores of the to-be-selected drivers.

[0161] The second obtaining unit is configured to obtain a goods-hunting list of the to-be-selected driver, and the goods sources in the goods-hunting list are sorted in descending order of the driver-goods matching degrees.

[0162] The second determining unit is configured to determine the first preset number of goods sources in the goods-hunting list as a driver-goods source pool, and determine the driver-goods matching degrees of the goods sources in the goods-hunting list as second driver-goods scores corresponding to the driver-goods source pool.

[0163] The third determining unit is configured to determine the to-be-matched goods sources, the to-be-selected drivers corresponding to the to-be-matched goods sources and the first driver-goods scores, the driver-goods source pool and the second driver-goods scores corresponding to the driver-goods source pool as driver-goods information corresponding to the to-be-matched goods sources.

[0164] In an embodiment, the driver-goods matching apparatus can further include an updating module configured to update the driver-goods source pool and the second driver-goods scores corresponding to the driver-goods source pool when detecting a goods-hunting operation of the to-be-selected driver.

[0165] In an embodiment, the driver-goods matching apparatus can further include a deleting module configured to delete a first goods source from the driver-goods source pool when detecting that the first goods source is offline and the first goods source is a cached goods source in the driver-goods source pool.

[0166] In an embodiment, the second determining module 220 can include:

[0167] The goods source list constructing unit is configured to construct the to-be-matched goods sources and the driver-goods source pool into a goods source list.

[0168] The third obtaining unit is configured to obtain, as a target goods source, a first utility value of the target goods source being selected and a second utility value of the goods source list not being selected in the goods source list, where the first utility value and the second utility value are used to represent the attractiveness of the target goods source.

[0169] The fourth determining unit is configured to determine a total expected utility value of all goods sources in the goods source list based on the first utility values and the fitting parameter.

[0170] The fifth determining unit is configured to determine, based on the fitting parameter, the total expected utility value and the second utility value, a first probability that the to-be-selected driver selects a goods source in the goods source list to make a deal and a second probability that the to-be-selected driver does not select a goods source in the goods source list to make a deal.

[0171] The sixth determining unit is configured to determine, for each target cargo source, a third probability that the target cargo source is selected based on the first utility value of the target cargo source.

[0172] The seventh determining unit is configured to determine, based on the first probability, the second probability and the third probability, a cargo source selection probability of the to-be-selected driver selecting each cargo source.

[0173] In a case where the target cargo source is a to-be-matched cargo source, the first utility value corresponding to the target cargo source is the first driver-cargo score; in a case where the target cargo source is a cargo source in the driver cargo source pool, the first utility value corresponding to the target cargo source is the second driver-cargo score; and the fitting parameter is determined by maximum log-likelihood estimation on historical driver behavior data.

[0174] In an embodiment, the seventh determining unit is specifically configured to determine, based on the first probability, the second probability and the third probability, the cargo source selection probability of the to-be-selected driver selecting each cargo source by using the following formula:

[0175]

[0176] wherein, denotes the cargo source selection probability of the jth to-be-selected driver selecting each cargo source, denotes the first probability, denotes the second probability, denotes the third probability, denotes the cargo source list, denotes the selection operation of the to-be-selected driver, denotes that no transaction is selected.

[0177] The driver-cargo matching apparatus provided in the embodiments of the present application has similar implementation principles and beneficial effects to the driver-cargo matching method provided in the above embodiments, and thus will not be described herein again.

[0178] The apparatus embodiments described above are merely schematic, wherein the units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e., can be located in one place or distributed on multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0179] The embodiments of the present application further provide an electronic device including a memory and a processor, the memory storing a computer program capable of running on the processor, and the processor implements the steps of the driver-cargo matching method of any of the above method embodiments when executing the computer program, and thus will not be described herein again.

[0180] ​Based on the cargo matching method described in any of the above embodiments, the embodiments of the present application further provide a computer readable storage medium, for example, a non-transitory computer readable storage medium can be a Read Only Memory (ROM), a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc. The storage medium stores computer instructions for executing the cargo matching method described in any of the above embodiments, which will not be described here.

[0181] Those skilled in the art can understand that all or part of the steps of the above-mentioned embodiments can be completed by hardware, or by program instructing relevant hardware to complete, and the program can be stored in a computer readable storage medium. The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0182] Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the application disclosed herein. The specification and examples are exemplary only, with the true scope and spirit of the application being indicated by the claims.

Claims

1. A method of matching a shipment, characterized by, The method comprises the following steps: determining the driver-goods information corresponding to the to-be-matched goods source, wherein the driver-goods information comprises the to-be-matched goods source, the to-be-selected driver corresponding to the to-be-matched goods source, the first driver-goods score, the driver-goods pool, and the second driver-goods score corresponding to the driver-goods pool; determining the goods source selection probability of the to-be-selected driver selecting each goods source based on the to-be-matched goods source, the first driver-goods score, the driver-goods pool, the second driver-goods score, and the historical transaction order of the to-be-selected driver; determining the target driver matched with the to-be-matched goods source from the to-be-selected driver based on the driver-goods information and the goods source selection probability, so that the sum of the transaction probabilities of the to-be-matched goods source and all goods sources in the driver-goods pool is maximum. The method comprises the following steps: obtaining the goods-to-car recall sorting information of the to-be-matched goods source, wherein the drivers in the goods-to-car recall sorting information are sorted in descending order of matching degree; determining the first preset number of drivers in the goods-to-car recall sorting information as the to-be-selected driver corresponding to the to-be-matched goods source, and determining the matching degree corresponding to each to-be-selected driver in the goods-to-car recall sorting information as the first driver-goods score of each to-be-selected driver; obtaining the goods finding list of the to-be-selected driver, wherein the goods sources in the goods finding list are sorted in descending order of driver-goods matching degree; determining the first second preset number of goods sources in the goods finding list as the driver-goods pool, and determining the driver-goods matching degree corresponding to each goods source in the goods finding list as the second driver-goods score corresponding to the driver-goods pool; determining the to-be-matched goods source, the to-be-selected driver corresponding to the to-be-matched goods source, the first driver-goods score, the driver-goods pool, and the second driver-goods score corresponding to the driver-goods pool as the driver-goods information corresponding to the to-be-matched goods source.

2. The load matching method of claim 1, wherein, The method comprises the following steps: based on the driver-goods information and the goods source selection probability, constructing a target function with the sum of the transaction probabilities of the to-be-matched goods source and all goods sources in the driver-goods pool being maximum as the target; taking the matching of the to-be-matched goods source to all drivers in the to-be-selected driver as the initial solution of the target function, subtracting the driver bringing the maximum gain value to the target function in each round of traversal based on the traversal pruning principle until no gain of the target function is generated, and determining the finally retained driver as the target driver matched with the to-be-matched goods source.

3. The load matching method of claim 2, wherein, The method comprises the following steps: taking the matching of the to-be-matched goods source to all drivers in the to-be-selected driver as the initial solution of the target function, subtracting the driver bringing the maximum gain value to the target function in each round of traversal based on the traversal pruning principle until no gain of the target function is generated, and determining the finally retained driver as the target driver matched with the to-be-matched goods source. determining, as an initial solution of the target function, a solution of matching the to-be-matched freight source to all drivers in the to-be-selected drivers, determining a target function value of the target function based on the initial solution; traversing each solution in the initial solution, determining a gain value brought to the target function value when the each solution is pruned; determining a maximum gain value in the gain values, and pruning the solution corresponding to the maximum gain value from the initial solution to obtain an updated initial solution; taking the updated initial solution as the initial solution of the target function again until the generated gain values are all less than or equal to 0, determining that no gain of the target function is generated, and determining the driver corresponding to the finally reserved solution in the initial solution as the target driver to which the to-be-matched freight source is matched.

4. The load matching method of claim 2, wherein, The target function is represented as: ； wherein P i represents the transaction probability of the i-th source, represents the function value of the target function, D represents the set of the to-be-selected drivers, C represents the set of the to-be-matched sources, represents the source selection probability of the j-th driver selecting the i-th source, represents the representation quantity of whether the i-th source is matched to the j-th driver; when represents that the i-th source is not matched to the j-th driver, and when represents that the i-th source is matched to the j-th driver.

5. The load dispatching method according to any one of claims 1 to 4, characterized by, The freight-driver matching method further includes: in the case of detecting the freight searching operation of the to-be-selected driver, updating the driver freight source pool and the second freight-driver score corresponding to the driver freight source pool; in the case of detecting that the first freight source is offline and the first freight source is the cached freight source in the driver freight source pool, deleting the first freight source from the driver freight source pool.

6. The load dispatching method according to any one of claims 1 to 4, characterized by, The determination of the freight source selection probability of the to-be-selected driver selecting each freight source based on the to-be-matched freight source, the first freight-driver score, the driver freight source pool, the second freight-driver score, and the historical transaction order of the to-be-selected driver includes: constructing the to-be-matched freight source and the driver freight source pool into a freight source list; taking each freight source in the freight source list as a target freight source, obtaining a first utility value of the target freight source being selected and a second utility value of the freight source list not being selected; the first utility value and the second utility value are used to represent the attractiveness of the target freight source; determining a total expected utility value of all freight sources in the freight source list based on each first utility value and a fitting parameter; based on the fitting parameter, the total expected utility value, and the second utility value, determining a first probability of the to-be-selected driver selecting a freight source in the freight source list for transaction and a second probability of the to-be-selected driver not selecting a freight source in the freight source list for transaction; for each target freight source, determining a third probability of the target freight source being selected based on the first utility value of the target freight source; determining the freight source selection probability of the to-be-selected driver selecting each freight source based on the first probability, the second probability, and the third probability; wherein, in the case of the target freight source being the to-be-matched freight source, the first utility value corresponding to the target freight source is the first freight-driver score; in the case of the target freight source being a freight source in the driver freight source pool, the first utility value corresponding to the target freight source is the second freight-driver score; and the fitting parameter is determined by maximum log-likelihood estimation on historical driver behavior data.

7. The load matching method of claim 6, wherein, The determination of the freight source selection probability of the to-be-selected driver selecting each freight source based on the first probability, the second probability, and the third probability includes: Based on the first probability, the second probability and the third probability, the cargo source selection probability of the to-be-selected driver selecting each cargo source is determined by using the following formula: ； wherein, denotes the probability of the jth driver selecting each of the sources, denotes the first probability, denotes the second probability, denotes the third probability, denotes the source list, and o denotes the selection operation of the driver to be selected, denotes that no transaction is selected.

8. A goods matching device, characterized by, Comprise: The first determining module is used for determining the driver-cargo information corresponding to the to-be-matched cargo source, wherein the driver-cargo information comprises the to-be-matched cargo source, the to-be-selected driver corresponding to the to-be-matched cargo source and the first driver-cargo score, the driver-cargo pool and the second driver-cargo score corresponding to the driver-cargo pool; The second determining module is used for determining the cargo source selection probability of the to-be-selected driver selecting each cargo source based on the to-be-matched cargo source, the first driver-cargo score, the driver-cargo pool, the second driver-cargo score and the historical transaction order of the to-be-selected driver; The third determining module is used for determining the target driver matched with the to-be-matched cargo source from the to-be-selected driver based on the driver-cargo information and the cargo source selection probability, so as to maximize the sum of the transaction probabilities of the to-be-matched cargo source and all cargo sources in the driver-cargo pool; The first determining module comprises: The first obtaining unit is used for obtaining the car-finding-cargo recall sorting information of the to-be-matched cargo source, wherein the drivers in the car-finding-cargo recall sorting information are sorted in descending order of matching degree; The first determining unit is used for determining the first preset number of drivers in the car-finding-cargo recall sorting information as the to-be-selected driver corresponding to the to-be-matched cargo source, and determining the matching degree corresponding to each to-be-selected driver in the car-finding-cargo recall sorting information as the first driver-cargo score of each to-be-selected driver; The second obtaining unit is used for obtaining the cargo-finding list of the to-be-selected driver, wherein the cargo sources in the cargo-finding list are sorted in descending order of driver-cargo matching degree; The second determining unit is used for determining the second preset number of cargo sources in the cargo-finding list as the driver-cargo pool, and determining the driver-cargo matching degree corresponding to each cargo source in the cargo-finding list as the second driver-cargo score corresponding to the driver-cargo pool; The third determining unit is used for determining the to-be-matched cargo source, the to-be-selected driver corresponding to the to-be-matched cargo source and the first driver-cargo score, the driver-cargo pool and the second driver-cargo score corresponding to the driver-cargo pool as the driver-cargo information corresponding to the to-be-matched cargo source.

9. An electronic device comprising a memory and a processor, said memory storing a computer program operable on said processor, characterized in that, The processor executes the computer program to realize the steps of the driver-cargo matching method in any one of claims 1 to 7.

Images