Method, device, server and storage medium for determining target operating force

Abstract

The application provides a method and device for determining target transport capacity, a server and a storage medium, and belongs to the technical field of computers. The method comprises the following steps: determining the ratio of the timeliness of the transport capacity of the target transport capacity type and the timeliness of the transport capacity of the reference transport capacity type as the timeliness ratio corresponding to the first weight group when the first weight group is applied; obtaining the delivery quality parameter reached when the first weight group is applied; determining, from a plurality of second weight groups, a target weight group that meets the timeliness ratio constraint condition and has the largest corresponding delivery quality parameter based on the timeliness ratios corresponding to the plurality of first weight groups and the delivery quality parameters; determining, based on the target weight group, the allocation parameters of the transport order for a plurality of transport capacities when a distribution request is received; and determining, from a plurality of transport capacities, a target transport capacity that meets the allocation condition of the allocation parameters. The application integrates the constraints of timeliness ratio and delivery quality in the process of transport capacity allocation, thereby ensuring the delivery quality and improving the stability of the transport capacity of the target transport capacity type.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, server and storage medium for determining target capacity. Background Technology

[0002] With the continuous development of computer technology, users can order food, shop online, or schedule delivery services through terminals. The delivery platform generates a waybill for the user based on the user's ordering, shopping, or scheduled delivery behavior, and then dispatches appropriate transportation capacity to deliver the items to the user based on the waybill.

[0003] In the context of on-demand delivery, delivery platforms typically rely on various types of transportation capacity to provide delivery services, in order to flexibly respond to changes in transportation capacity demand at different times. Some of these transportation capacities have flexible working hours, while others have relatively stable working hours and are able to cope with complex delivery scenarios such as severe weather and insufficient transportation capacity.

[0004] Currently, delivery platforms determine which delivery capacity offers the best route alignment and lowest risk of delay based on indicators such as the likelihood of delivery to the order and the risk of delay. However, when the total number of orders is relatively small, but the number of delivery personnel with flexible operating hours is large, it can easily lead to a decrease in the number of orders allocated to delivery personnel with relatively stable operating hours. This negatively impacts the delivery experience, causes capacity loss, and, in the long run, adversely affects delivery quality. Therefore, how to rationally allocate various types of delivery capacity to maintain the stability of delivery personnel with relatively stable operating hours while ensuring delivery quality is a problem that urgently needs to be solved. Summary of the Invention

[0005] This application provides a method, apparatus, server, and storage medium for determining target transportation capacity, which can maintain the stability of transportation capacity with relatively stable attendance times while ensuring delivery quality. The technical solution is as follows:

[0006] According to one aspect of the embodiments of this application, a method for determining target transportation capacity is provided, the method comprising:

[0007] For any first weighting reorganization in a series of first weighting reorganizations applied to capacity allocation in the past, the ratio of the timeliness of the capacity of the target capacity type when the first weighting reorganization is applied to the timeliness of the capacity of the reference capacity type when the first weighting reorganization is applied is determined as the timeliness ratio corresponding to the first weighting reorganization, and the delivery quality parameters achieved when the first weighting reorganization is applied are obtained.

[0008] Based on the timeliness ratio and delivery quality parameters corresponding to the multiple first weighted reorganizations, from the multiple second weighted reorganizations that have not yet been applied to capacity allocation, a target weighted reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter is determined. The timeliness ratio constraint is that the timeliness ratio is greater than or equal to the timeliness ratio threshold.

[0009] Upon receiving an allocation request for any waybill, based on the target weighting reorganization, the allocation parameters of the waybill for multiple transport capacities are determined; from the multiple transport capacities, a target transport capacity whose allocation parameters satisfy the allocation conditions is determined, and the allocation parameters of any transport capacity are used to represent the overall matching degree between the waybill and the transport capacity;

[0010] Among them, timeliness is used to represent the number of waybills delivered per unit of capacity per unit of time; any weighting group includes a first weight and a second weight applied when allocating capacity, the first weight is used to represent the importance of the degree of matching between capacity and waybill in capacity allocation, and the second weight is used to represent the importance of capacity type in capacity allocation.

[0011] In one possible implementation, determining the target weighted reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from multiple second weighted reorganizations that have not yet been applied to capacity allocation, based on the timeliness ratio and delivery quality parameters corresponding to the multiple first weighted reorganizations, includes:

[0012] Based on the plurality of first weighted reorganizations, the delivery quality parameters corresponding to the plurality of first weighted reorganizations, and the timeliness ratios corresponding to the plurality of first weighted reorganizations, a first distribution of the delivery quality parameters corresponding to the plurality of second weighted reorganizations and a second distribution of the timeliness ratios corresponding to the plurality of second weighted reorganizations are obtained;

[0013] Based on the first distribution and the second distribution corresponding to the multiple second weighting reorganizations, the target weighting reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter is determined from the multiple second weighting reorganizations.

[0014] The first distribution includes multiple delivery quality parameters corresponding to each second weighting reorganization and the probability of achieving each delivery quality parameter under the premise of applying each second weighting reorganization;

[0015] The second distribution includes multiple time efficiency ratios corresponding to each second weighting reorganization and the probability of achieving each time efficiency ratio under the premise of applying each second weighting reorganization.

[0016] In another possible implementation, obtaining a first distribution of the delivery quality parameters corresponding to the multiple second weighted reassemblies and a second distribution of the timeliness ratios corresponding to the multiple second weighted reassemblies based on the multiple first weighted reassemblies, the delivery quality parameters corresponding to the multiple first weighted reassemblies, and the timeliness ratios corresponding to the multiple first weighted reassemblies includes:

[0017] Based on the multiple first weighted reorganizations and the delivery quality parameters corresponding to the multiple first weighted reorganizations, Gaussian process regression is performed to obtain the first distribution of the delivery quality parameters corresponding to the multiple second weighted reorganizations.

[0018] Based on the multiple first-weighted reorganizations and the time efficiency ratios corresponding to the multiple first-weighted reorganizations, a Gaussian process regression is performed to obtain the second distribution of the time efficiency ratios corresponding to the multiple second-weighted reorganizations.

[0019] In another possible implementation, determining the target weighting arrangement that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from the plurality of second weighting arrangements based on the first distribution and the second distribution of the plurality of second weighting arrangements includes:

[0020] Based on the first distribution of the multiple second weighting reorganizations, the second distribution of the multiple second weighting reorganizations, and the reference delivery quality parameters, the weight selection reference values ​​for the multiple second weighting reorganizations are determined.

[0021] From the plurality of second weighting reorganizations, determine the target weighting reorganization with the largest reference value for the corresponding weight selection;

[0022] The reference delivery quality parameter is the largest delivery quality parameter among the multiple first weighted groups, and the reference value for weight selection is a comprehensive index that integrates the timeliness ratio constraint and the delivery quality parameter.

[0023] In another possible implementation, determining the weight selection reference value for the plurality of second weight reassemblies based on the first distribution corresponding to the plurality of second weight reassemblies, the second distribution corresponding to the plurality of second weight reassemblies, and reference delivery quality parameters includes:

[0024] For any second weighting reorganization among the plurality of second weighting reorganizations, based on the first distribution corresponding to the plurality of second weighting reorganizations and the reference delivery quality parameter, the expected delivery quality corresponding to the second weighting reorganization is determined. The expected delivery quality is used to represent the degree of improvement of the delivery quality parameter achieved under the premise of applying the second weighting reorganization compared to the reference delivery quality parameter.

[0025] Based on the second distribution corresponding to the multiple second weighting reorganizations, the probability of condition satisfaction corresponding to the second weighting reorganization is determined. The probability of condition satisfaction refers to the probability that the time efficiency ratio achieved under the premise of applying the second weighting reorganization satisfies the time efficiency ratio constraint condition.

[0026] The product of the expected delivery quality and the probability of the condition being met is determined as the reference value for selecting the weights corresponding to the second weighted reassembly.

[0027] In another possible implementation, determining the expected delivery quality corresponding to the second weighting based on the first distribution of the plurality of second weightings and the reference delivery quality parameters includes:

[0028] Based on the first distribution corresponding to the multiple second weighting reorganizations, multiple target delivery quality parameters that are greater than the reference delivery quality parameter are obtained from the multiple delivery quality parameters corresponding to the second weighting reorganizations, and the probabilities corresponding to the multiple target delivery quality parameters are obtained. The probability corresponding to any target delivery quality parameter refers to the probability of achieving any target delivery quality parameter under the premise of applying the second weighting reorganization.

[0029] For any one of the plurality of target delivery quality parameters, determine the difference between the target delivery quality parameter and the reference delivery quality parameter; determine the product of the difference corresponding to the target delivery quality parameter and the probability corresponding to the target delivery quality parameter;

[0030] The mean of the products corresponding to the multiple target delivery quality parameters is determined as the configuration quality expectation corresponding to the second weighted reassembly.

[0031] In another possible implementation, determining the probability of satisfying the condition corresponding to the second weighting reorganization based on the second distribution corresponding to the plurality of second weighting reorganizations includes:

[0032] Based on the second distribution corresponding to the multiple second weighting reorganizations, multiple target time efficiency ratios that are greater than the time efficiency ratio threshold are determined from the multiple time efficiency ratios corresponding to the second weighting reorganizations;

[0033] The sum of the probabilities corresponding to the multiple target timeliness ratios is determined as the probability of the condition being met;

[0034] The probability corresponding to any target timeliness ratio refers to the probability of achieving any target timeliness ratio under the premise of applying the second weighting reorganization.

[0035] In another possible implementation, the method further includes:

[0036] If the duration for which the target weight reorganization is applied to capacity allocation reaches the target duration and the feedback adjustment conditions are met, the target weight reorganization is taken as the first weight reorganization. The following steps are performed: for any one of the multiple first weight reorganizations applied to capacity allocation in the past, the ratio of the timeliness of the capacity of the target capacity type when the first weight reorganization is applied to the timeliness of the capacity of the reference capacity type when the first weight reorganization is applied is determined as the timeliness ratio corresponding to the first weight reorganization; and the delivery quality parameters achieved when the first weight reorganization is applied are obtained.

[0037] In another possible implementation, the feedback adjustment condition includes at least one of a first feedback adjustment condition and a second feedback adjustment condition;

[0038] The first feedback adjustment condition refers to the fact that the number of feedback adjustments has not reached the upper limit;

[0039] The second feedback adjustment condition refers to the delivery quality parameter achieved when the target weighting is applied being greater than the reference delivery quality parameter, where the reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to the plurality of first weightings.

[0040] In another possible implementation, determining the allocation parameters of the waybill for multiple capacities based on the target weight reorganization includes:

[0041] For any one of the multiple transport capacities, obtain the delivery matching parameters of the transport capacity, and obtain the level parameters corresponding to the transport capacity type to which the transport capacity belongs. The delivery matching parameters are used to represent the degree of delivery matching between the transport capacity and the waybill.

[0042] Determine the first product of the delivery matching parameters of the transport capacity and the first weight in the target weighting;

[0043] Determine the second product of the capacity level parameter and the second weight in the target weighting;

[0044] The sum of the first product and the second product is determined as the allocation parameter of the transport capacity.

[0045] According to another aspect of the embodiments of this application, an apparatus for determining target transport capacity is provided, the apparatus comprising:

[0046] The timeliness ratio determination module is used to determine the timeliness ratio corresponding to the first weighted reorganization as the ratio of the timeliness of the target capacity type when the first weighted reorganization is applied to the timeliness of the reference capacity type when the first weighted reorganization is applied for any first weighted reorganization in a series of first weighted reorganizations applied in the past.

[0047] The quality parameter acquisition module is used to acquire the delivery quality parameters achieved when the first weighting is applied.

[0048] The target weight reorganization determination module is used to determine, based on the timeliness ratio and delivery quality parameters corresponding to the multiple first weight reorganizations, the target weight reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from the multiple second weight reorganizations that have not yet been applied to capacity allocation. The timeliness ratio constraint is that the timeliness ratio is greater than or equal to the timeliness ratio threshold.

[0049] The allocation parameter determination module is used to determine the allocation parameters of the waybill for multiple transport capacities based on the target weighting reorganization when receiving an allocation request for any waybill. The allocation parameter of any transport capacity is used to represent the overall matching degree between the waybill and the transport capacity.

[0050] The target capacity determination module is used to determine the target capacity whose allocation parameters satisfy the allocation conditions from the plurality of capacity;

[0051] Among them, timeliness is used to represent the number of waybills delivered per unit of capacity per unit of time; any weighting group includes a first weight and a second weight applied when allocating capacity, the first weight is used to represent the importance of the degree of matching between capacity and waybill in capacity allocation, and the second weight is used to represent the importance of capacity type in capacity allocation.

[0052] In one possible implementation, the target weight reorganization determination module includes:

[0053] The distribution acquisition unit is used to acquire, based on the plurality of first weighted reassemblies, the delivery quality parameters corresponding to the plurality of first weighted reassemblies, and the timeliness ratios corresponding to the plurality of first weighted reassemblies, the first distribution of delivery quality parameters corresponding to the plurality of second weighted reassemblies and the second distribution of timeliness ratios corresponding to the plurality of second weighted reassemblies;

[0054] The target weight reorganization determination unit is used to determine, based on the first distribution and the second distribution corresponding to the plurality of second weight reorganizations, the target weight reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from the plurality of second weight reorganizations.

[0055] The first distribution includes multiple delivery quality parameters corresponding to each second weighting reorganization and the probability of achieving each delivery quality parameter under the premise of applying each second weighting reorganization;

[0056] The second distribution includes multiple time efficiency ratios corresponding to each second weighting reorganization and the probability of achieving each time efficiency ratio under the premise of applying each second weighting reorganization.

[0057] In another possible implementation, the distribution acquisition unit is used for:

[0058] Based on the multiple first weighted reorganizations and the delivery quality parameters corresponding to the multiple first weighted reorganizations, Gaussian process regression is performed to obtain the first distribution of the delivery quality parameters corresponding to the multiple second weighted reorganizations.

[0059] Based on the multiple first-weighted reorganizations and the time efficiency ratios corresponding to the multiple first-weighted reorganizations, a Gaussian process regression is performed to obtain the second distribution of the time efficiency ratios corresponding to the multiple second-weighted reorganizations.

[0060] In another possible implementation, the target weight reorganization determining unit includes:

[0061] The reference value determination subunit is used to determine the weight selection reference value corresponding to the multiple second weight reassemblies based on the first distribution corresponding to the multiple second weight reassemblies, the second distribution corresponding to the multiple second weight reassemblies, and the reference delivery quality parameters;

[0062] The target weight reorganization determination subunit is used to determine the target weight reorganization with the largest corresponding weight selection reference value from the plurality of second weight reorganizations;

[0063] The reference delivery quality parameter is the largest delivery quality parameter among the multiple first weighted groups, and the reference value for weight selection is a comprehensive index that integrates the timeliness ratio constraint and the delivery quality parameter.

[0064] In another possible implementation, the reference value determines the sub-unit for:

[0065] For any second weighting reorganization among the plurality of second weighting reorganizations, based on the first distribution corresponding to the plurality of second weighting reorganizations and the reference delivery quality parameter, the expected delivery quality corresponding to the second weighting reorganization is determined. The expected delivery quality is used to represent the degree of improvement of the delivery quality parameter achieved under the premise of applying the second weighting reorganization compared to the reference delivery quality parameter.

[0066] Based on the second distribution corresponding to the multiple second weighting reorganizations, the probability of condition satisfaction corresponding to the second weighting reorganization is determined. The probability of condition satisfaction refers to the probability that the time efficiency ratio achieved under the premise of applying the second weighting reorganization satisfies the time efficiency ratio constraint condition.

[0067] The product of the expected delivery quality and the probability of the condition being met is determined as the reference value for selecting the weights corresponding to the second weighted reassembly.

[0068] In another possible implementation, the reference value determines the sub-unit for:

[0069] Based on the first distribution corresponding to the multiple second weighting reorganizations, multiple target delivery quality parameters that are greater than the reference delivery quality parameter are obtained from the multiple delivery quality parameters corresponding to the second weighting reorganizations, and the probabilities corresponding to the multiple target delivery quality parameters are obtained. The probability corresponding to any target delivery quality parameter refers to the probability of achieving any target delivery quality parameter under the premise of applying the second weighting reorganization.

[0070] For any one of the plurality of target delivery quality parameters, determine the difference between the target delivery quality parameter and the reference delivery quality parameter; determine the product of the difference corresponding to the target delivery quality parameter and the probability corresponding to the target delivery quality parameter;

[0071] The mean of the products corresponding to the multiple target delivery quality parameters is determined as the configuration quality expectation corresponding to the second weighted reassembly.

[0072] In another possible implementation, the reference value determines the sub-unit for:

[0073] Based on the second distribution corresponding to the multiple second weighting reorganizations, multiple target time efficiency ratios that are greater than the time efficiency ratio threshold are determined from the multiple time efficiency ratios corresponding to the second weighting reorganizations;

[0074] The sum of the probabilities corresponding to the multiple target timeliness ratios is determined as the probability of the condition being met;

[0075] The probability corresponding to any target timeliness ratio refers to the probability of achieving any target timeliness ratio under the premise of applying the second weighting reorganization.

[0076] In another possible implementation, the device further includes:

[0077] The feedback adjustment unit is configured to, if the duration for which the target weight reorganization is applied to capacity allocation reaches the target duration and the feedback adjustment conditions are met, designate the target weight reorganization as the first weight reorganization; for any first weight reorganization among multiple first weight reorganizations historically applied to capacity allocation, determine the ratio of the timeliness of the capacity of the target capacity type when the first weight reorganization is applied to the timeliness of the capacity of the reference capacity type when the first weight reorganization is applied as the timeliness ratio corresponding to the first weight reorganization; and obtain the delivery quality parameters achieved when the first weight reorganization is applied.

[0078] In another possible implementation, the feedback adjustment condition includes at least one of a first feedback adjustment condition and a second feedback adjustment condition;

[0079] The first feedback adjustment condition refers to the fact that the number of feedback adjustments has not reached the upper limit;

[0080] The second feedback adjustment condition refers to the delivery quality parameter achieved when the target weighting is applied being greater than the reference delivery quality parameter, where the reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to the plurality of first weightings.

[0081] In another possible implementation, the allocation parameter determination module is used for:

[0082] For any one of the multiple transport capacities, obtain the delivery matching parameters of the transport capacity, and obtain the level parameters corresponding to the transport capacity type to which the transport capacity belongs. The delivery matching parameters are used to represent the degree of delivery matching between the transport capacity and the waybill.

[0083] Determine the first product of the delivery matching parameters of the transport capacity and the first weight in the target weighting;

[0084] Determine the second product of the capacity level parameter and the second weight in the target weighting;

[0085] The sum of the first product and the second product is determined as the allocation parameter of the transport capacity.

[0086] According to another aspect of the embodiments of this application, a server is provided, the server including a processor and a memory, the memory storing at least one piece of program code, the at least one piece of program code being loaded and executed by the processor to implement the method for determining target capacity as described in any of the possible implementations above.

[0087] According to another aspect of the embodiments of this application, a computer-readable storage medium is provided, wherein at least one piece of program code is stored in the computer-readable storage medium, the at least one piece of program code being loaded and executed by a processor to implement the method for determining target capacity as described in any of the above possible implementations.

[0088] According to another aspect of the embodiments of this application, a computer program product or computer program is provided, which includes computer program code stored in a computer-readable storage medium. The processor of a server reads the computer program code from the computer-readable storage medium and executes the computer program code, causing the server to perform the method for determining target capacity as described in any of the above possible implementations.

[0089] The technical solutions provided in this application have at least the following beneficial effects:

[0090] By setting timeliness ratio constraints, the timeliness of the target capacity type is made greater than that of the reference capacity type, thus maintaining the bias of waybills towards the target capacity type. Simultaneously, by setting a condition that maximizes the delivery quality parameter, delivery quality is guaranteed. Based on this, referencing the delivery quality parameters and timeliness ratios achieved when applying the first weighting reorganization in historical time periods, a target weighting reorganization that satisfies both the timeliness ratio constraint and the maximum delivery quality parameter is determined from multiple unapplied second weighting reorganizations. This target weighting reorganization is then applied for capacity allocation. This allocation process incorporates considerations of both delivery quality and the bias towards the target capacity type, ensuring delivery quality while maintaining the bias in waybill allocation towards the target capacity type. This reduces the loss of target capacity and improves its stability. Attached Figure Description

[0091] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0092] Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application;

[0093] Figure 2 This is a flowchart illustrating a method for determining target transportation capacity provided in an embodiment of this application;

[0094] Figure 3 This is a flowchart illustrating a method for determining target transportation capacity provided in an embodiment of this application;

[0095] Figure 4 This is a schematic diagram of a joint Gaussian distribution provided in an embodiment of this application;

[0096] Figure 5 This is a schematic diagram of a one-dimensional Gaussian distribution provided in an embodiment of this application;

[0097] Figure 6 This is a schematic diagram of a one-dimensional Gaussian distribution provided in an embodiment of this application;

[0098] Figure 7 This is a flowchart illustrating a method for determining target transportation capacity provided in an embodiment of this application;

[0099] Figure 8 This is a comparison chart of relative on-time performance provided in an embodiment of this application;

[0100] Figure 9 This is a comparison chart of the percentage of daily delivery orders exceeding 55 minutes, provided in an embodiment of this application.

[0101] Figure 10 This is a comparison chart of average delivery time per order provided in an embodiment of this application;

[0102] Figure 11 This is a block diagram of a device for determining target transport capacity provided in an embodiment of this application;

[0103] Figure 12 This is a block diagram of a server provided in an embodiment of this application. Detailed Implementation

[0104] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0105] The terms "first," "second," "third," and "fourth," etc., used in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0106] Figure 1 This is a schematic diagram of an implementation environment provided in an embodiment of this application. See also... Figure 1 The implementation environment includes server 101 and terminal 102.

[0107] Optionally, server 101 is a single server; or server 101 is a server cluster or distributed system composed of several servers; or server 101 is a cloud computing service center. Server 101 has the function of determining the target capacity for waybills.

[0108] Server 101 and terminal 102 are connected via a wireless or wired network. After determining the target capacity for a waybill, server 101 sends the delivery information for that waybill to the terminal 102 corresponding to the target capacity. Optionally, terminal 102 can be a smartphone, wearable device, tablet, laptop, or desktop computer, but is not limited to these. After receiving the delivery information sent by server 101, terminal 102 displays the delivery information and prompts the target capacity to deliver the goods based on the delivery information. Optionally, terminal 102 has a delivery client, which assists the target capacity in delivering the goods. By running the delivery client, terminal 102 receives the delivery information sent by server 101, displays the delivery information, and prompts the target capacity to deliver the goods based on the delivery information.

[0109] Figure 2 This is a flowchart illustrating a method for determining target transportation capacity provided in an embodiment of this application. The following is in conjunction with… Figure 2 A brief explanation of the method for determining the target capacity is provided below. Figure 2 The method for determining the target capacity is executed by the server and includes the following steps:

[0110] 201. For any first-weighted reorganization among multiple first-weighted reorganizations applied to capacity allocation in the past, the server determines the timeliness ratio corresponding to the first-weighted reorganization based on the timeliness of the capacity of the target capacity type when the first-weighted reorganization is applied and the timeliness of the capacity of the reference capacity type when the first-weighted reorganization is applied, and obtains the delivery quality parameters achieved when the first-weighted reorganization is applied.

[0111] In this context, "transportation capacity" refers to the transportation force that moves goods from their origin to their destination; in other words, it refers to the delivery personnel who transport goods from their origin to their destination. In the data stored on the server, transportation capacity is represented by a capacity identifier, with different capacity types corresponding to different identifiers. Furthermore, in the context of food delivery, transportation capacity can also be referred to as riders, delivery personnel, or couriers; in the context of express delivery, it can also be referred to as couriers or couriers.

[0112] The target capacity type is characterized by stable operating hours, enabling it to handle complex delivery scenarios such as holidays and inclement weather. Even in these challenging conditions, it can still accept orders and deliver goods to users. By relying on this target capacity type, the delivery platform can improve overall delivery quality and enhance the user experience. In some embodiments, the delivery platform ensures the stability of the target capacity type's operating hours by constraining these hours, provided that the capacity chooses to accept operating hour constraints. For example, the platform can constrain the target capacity type's operating hours by limiting the number of days and duration of operation per week. Alternatively, it can constrain the target capacity type's operating hours by limiting the number of days and duration of operation per week.

[0113] The transportation capacity referred to is that of part-time delivery personnel. Their attendance time is variable and their capacity is highly flexible. They can assist in delivery during peak hours and reduce delivery pressure, but they are difficult to cope with complex delivery scenarios. In complex delivery scenarios, they have a certain negative impact on delivery quality.

[0114] Capacity allocation refers to the process of allocating capacity to waybills, that is, the process of determining target capacity for waybills. The first weighting set is a weighting set that has been applied to capacity allocation within a historical time period. The weighting set involved in this application embodiment includes at least two weights, each weight in the weighting set representing the importance of different aspects related to capacity allocation. The value range of each weight in the weighting set is from 0 to 1, that is, each weight in the weighting set is a value greater than or equal to 0 and less than or equal to 1. Furthermore, the sum of the multiple weights included in the weighting set is 1. In some embodiments, any weighting set includes a first weight and a second weight applied during capacity allocation. The first weight represents the importance of the delivery matching degree between capacity and waybill to capacity allocation, and the second weight represents the importance of capacity type to capacity allocation.

[0115] Efficiency refers to the number of waybills delivered per unit of transport capacity per unit of time, that is, the number of waybills completed for delivery per unit of transport capacity within a unit of time. For example, if the unit of time is 1 hour, and a certain transport capacity completes delivery of 10 waybills in 2 hours, then the efficiency of that transport capacity is 10 / 2 = 5 waybills / hour. In the embodiments of this application, the transport capacity of the target transport capacity type is a collective term for multiple transport capacities of the target transport capacity type, and the efficiency of the transport capacity of the target transport capacity type refers to the average efficiency of the multiple transport capacities of the target transport capacity type, that is, the mean of the efficiency of the multiple transport capacities of the target transport capacity type. For example, if there are 3 multiple transport capacities of the target transport capacity type, and the efficiency of these 3 target transport capacity types is 5 waybills / hour, 6 waybills / hour, and 10 waybills / hour respectively, then the efficiency of the transport capacity of the target transport capacity type is (5+6+10) / 3 = 7 waybills / hour. Similarly, the efficiency of the transport capacity of the reference transport capacity type refers to the average efficiency of the multiple transport capacities of the reference transport capacity type.

[0116] 202. Based on the timeliness ratio and delivery quality parameters corresponding to multiple first weighted reorganizations, the server determines the target weighted reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from multiple second weighted reorganizations that have not yet been applied to capacity allocation. The timeliness ratio constraint is that the timeliness ratio is greater than or equal to the timeliness ratio threshold.

[0117] Because target capacity types offer advantages such as stable operating hours and the ability to handle complex delivery scenarios, they are crucial for ensuring stable delivery capacity and improving overall delivery quality. Therefore, to maintain a steady improvement in overall delivery quality, the stability of target capacity types should be maintained. For example, this can be achieved by increasing the allocation of orders to target capacity types, increasing the number of orders assigned to them, and ensuring that the delivery time of target capacity types is longer than that of reference capacity types. This incentivizes existing target capacity types to maintain stable operating hours and attracts more delivery personnel to join target capacity types, thereby reducing the loss of target capacity and maintaining their stability.

[0118] However, the allocation of waybills to the target capacity type should not be too large, otherwise it will result in too many waybills being allocated to the target capacity type, putting too much pressure on the target capacity type's delivery, causing some waybills to be delivered late, and adversely affecting the delivery quality.

[0119] Therefore, in order to ensure delivery quality while maintaining the stability of the target capacity type, capacity allocation should comprehensively consider both delivery quality and timeliness. In this embodiment, the weighting system used for capacity allocation includes weights representing the importance of the delivery matching degree between waybills and capacity, and weights representing the importance of capacity type. This weighting system influences capacity allocation from both the delivery matching degree and capacity type perspectives, thereby affecting delivery quality and timeliness under that allocation. Based on this, by setting conditions for maximizing delivery quality parameters and timeliness constraints, a target weighting system that meets these conditions can be found. Applying this target weighting system can ensure delivery quality while maintaining the stability of the target capacity type.

[0120] The first weighted reassembly refers to weighted reassemblies that have been applied to capacity allocation within a historical time period. The delivery quality parameters and timeliness ratios corresponding to the first weighted reassembly refer to the actual delivery quality parameters and timeliness ratios achieved when the first weighted reassembly was applied, and these parameters and ratios are known. The second weighted reassembly refers to weighted reassemblies that have not yet been applied to capacity allocation. The delivery quality parameters and timeliness ratios corresponding to the second weighted reassembly refer to the predicted delivery quality parameters and timeliness ratios that can be achieved under the premise of applying the second weighted reassembly, and these parameters and ratios are unknown. Therefore, based on the known timeliness ratios and delivery quality parameters corresponding to multiple first weighted reassemblies, the timeliness ratios and delivery quality parameters corresponding to multiple second weighted reassemblies are predicted. Then, from the multiple second weighted reassemblies, the target weighted reassembly with the largest delivery quality parameter that satisfies the timeliness ratio constraint is determined. The target weighted reassembly is the second weighted reassembly with the largest delivery quality parameter among the multiple second weighted reassemblies that satisfy the timeliness ratio constraint.

[0121] In some embodiments, multiple first weighting sets are weighting sets applied to capacity allocation within a historical time period relatively close to the current time. The timeliness ratio and delivery quality parameters corresponding to these multiple first weighting sets have high reference value, which helps to determine the target weighting set that is more suitable for the current application scenario. That is, the difference between the application time of the above-mentioned first weighting set and the current time is less than a time threshold. This time threshold can be flexibly configured, for example, the time threshold is 1 month or 3 months, etc., and this application embodiment does not limit it in this way.

[0122] 203. When the server receives an allocation request for any waybill, it determines the allocation parameters of the waybill for multiple transport capacities based on the target weight reorganization; it then determines the target transport capacity whose allocation parameters meet the allocation conditions from among the multiple transport capacities, wherein the allocation parameters of any transport capacity are used to represent the overall matching degree between the waybill and the transport capacity.

[0123] The target weighting process includes a first weight representing the importance of the delivery matching degree between the transport capacity and the waybill to the transport capacity allocation, and a second weight representing the importance of the transport capacity type to the transport capacity allocation. For a waybill and any one of multiple transport capacities, based on the first weight in the target weighting process, the parameter representing the delivery matching degree between the transport capacity and the waybill, the second weight, and the parameter representing the transport capacity type to which the transport capacity belongs, the allocation parameters for the waybill to the transport capacity are determined, thereby obtaining the allocation parameters for multiple transport capacities. Then, based on the allocation parameters of multiple transport capacities, a target transport capacity whose allocation parameters meet the allocation conditions is determined from the multiple transport capacities, and the waybill is allocated to the target transport capacity, which then delivers the goods indicated by the waybill.

[0124] The technical solution provided in this application sets a timeliness ratio constraint to ensure that the timeliness of the target capacity type is greater than that of the reference capacity type, thus maintaining the bias of waybills towards the target capacity type. Simultaneously, it sets a condition for maximizing the delivery quality parameter to guarantee delivery quality. Based on this, referencing the delivery quality parameters and timeliness ratios achieved when applying the first weighting reorganization in historical time periods, a target weighting reorganization that satisfies both the timeliness ratio constraint and the maximum delivery quality parameter is determined from multiple unapplied second weighting reorganizations. This target weighting reorganization is then applied for capacity allocation. This capacity allocation process incorporates considerations of both delivery quality and the bias towards the target capacity type, ensuring delivery quality while maintaining the bias of waybill allocation towards the target capacity type. This reduces the loss of target capacity type capacity and improves the stability of the target capacity type.

[0125] Figure 3 This is a flowchart illustrating a method for determining target transportation capacity provided in an embodiment of this application. The following is in conjunction with… Figure 3 For a detailed explanation of the method for determining the target capacity, please refer to [link / reference]. Figure 3 The method for determining the target capacity is executed by the server and includes the following steps:

[0126] 301. For any first-weighted reorganization among multiple first-weighted reorganizations applied to capacity allocation in the past, the server will determine the timeliness ratio corresponding to the first-weighted reorganization as the ratio of the timeliness of the capacity of the target capacity type when the first-weighted reorganization is applied to the timeliness of the capacity of the reference capacity type when the first-weighted reorganization is applied.

[0127] In this context, the timeliness of the target capacity type when applying the first weighting reorganization refers to the average timeliness of multiple capacity types of the target capacity participating in the delivery of goods within the effective duration of the first weighting group. The timeliness of the reference capacity type when applying the first weighting reorganization refers to the average timeliness of multiple capacity types of the reference capacity participating in the delivery of goods within the effective duration of the first weighting group. The effective duration refers to the duration applied to capacity allocation after the weighting reorganization takes effect. The effective duration of the weighting group can be flexibly configured; for example, the effective duration of a weighting group can be 1 day, 3 days, or 1 week, etc., and this embodiment does not impose any limitations on this.

[0128] In one example, the first weight group is effective for one day. During this day, there are three target capacity types participating in the delivery of goods, with delivery times of 5 orders / hour, 10 orders / hour, and 9 orders / hour, respectively. Therefore, the delivery time of the target capacity types when applying the first weight group is (5+10+9) / 3 = 8 orders / hour. If, during the same day, there are two target capacity types participating in the delivery of goods, with delivery times of 2 orders / hour and 1 order / hour, respectively, the delivery time of the reference capacity types when applying the first weight group is (2+1) / 2 = 1.5 orders / hour. The delivery time ratio corresponding to this first weight group is 8 / 1.5.

[0129] The delivery time of any transport capacity within the effective period of the first weighted group is defined as the ratio of the number of delivery orders completed by that transport capacity within the effective period of the first weighted group to the total working hours of that transport capacity within the effective period of the first weighted group. The working hours of a transport capacity within the effective period of the first weighted group refer to the duration of time that transport capacity participates in goods delivery work within the effective period of the first weighted group. For example, if the effective period of the first weighted group is 1 day, and a transport capacity works 8 hours within that day and completes 40 delivery orders, then the delivery time of that transport capacity within the effective period of the first weighted group is 40 / 8 = 5 orders / hour.

[0130] It should be noted that, for ease of understanding, the above example uses only a small amount of transportation capacity involved in the delivery of goods. In actual application scenarios, the number of transportation capacities involved in the delivery of goods when applying the first-weighted reorganization may be more or less, such as tens of thousands, hundreds of thousands or millions, but the method of determining the timeliness ratio corresponding to the first-weighted reorganization is the same as the above example, and will not be repeated here.

[0131] 302. The server obtains the delivery quality parameters achieved during the application's first-order reorganization.

[0132] The delivery quality parameter is used to represent the quality of delivery. The delivery quality parameter achieved when the first weighting group is applied represents the delivery quality achieved within the effective period of the first weighting group. Additionally, the delivery quality parameter also reflects the user experience of the delivery service; therefore, it can also be called an experience indicator.

[0133] In one possible implementation, the delivery quality parameter is a comprehensive delivery quality parameter determined based on delivery quality indicators across multiple dimensions. The steps for the server to obtain the delivery quality parameter achieved during the first weighted reorganization of the application include: the server obtaining delivery quality indicators across multiple dimensions during the first weighted reorganization; obtaining the influence weights corresponding to the delivery quality indicators across multiple dimensions, where each delivery quality indicator in a dimension corresponds to an influence weight, and the influence weight of each dimension's delivery quality indicator represents its importance; and the server performing a weighted summation of the delivery quality indicators across multiple dimensions during the first weighted reorganization based on the influence weights corresponding to these indicators, to obtain the delivery quality parameter achieved during the first weighted reorganization. That is, the server determines the delivery quality parameter achieved during the first weighted reorganization based on the following formula:

[0134] Formula 1:

[0135] Where, obj represents the delivery quality parameter; n represents the number of delivery quality indicators across multiple dimensions, where n is an integer greater than or equal to 1; w i f represents the influence weight of the delivery quality indicator in the i-th dimension; i is an integer greater than or equal to 1 and less than or equal to n; i (α,β) represents the delivery quality index achieved in the i-th dimension when the first weighting is applied; α is the first weight in the first weighting; β is the second weight in the first weighting.

[0136] Optionally, in on-demand delivery scenarios, such as food delivery, the delivery quality indicators across multiple dimensions include at least two of the following: average delivery time per order, relative on-time rate, order completion rate, 55-minute delay rate, and ETA+15 delay rate. Of course, depending on the specific application scenario, the delivery quality indicators across multiple dimensions may also include other indicators used to represent delivery quality, such as item integrity rate; this embodiment of the application does not impose any limitations on this.

[0137] The average delivery time per order refers to the average delivery time of multiple orders completed within the effective period of the first weighted group. The relative on-time rate refers to the percentage of orders completed within the effective period of the first weighted group whose actual delivery time is before the estimated delivery time. Order completion means that the order is completed by the initially allocated capacity and is not transferred to other capacity. The order completion rate is the percentage of multiple orders generated within the effective period of the first weighted group that are completed by the initially allocated capacity. The 55-minute delay rate is the percentage of multiple orders completed within the effective period of the first weighted group whose delivery time is longer than 55 minutes. ETA stands for Estimated Time of Arrival. The ETA+15 overtime rate refers to the percentage of multiple orders that are delivered within the effective period of the first weight group, where the actual delivery time is after the estimated delivery time, and the difference between the actual delivery time and the estimated delivery time is greater than 15 minutes.

[0138] It's important to note that delivery quality parameters are positively correlated with delivery quality; that is, a higher value indicates better delivery quality, and a lower value indicates worse delivery quality. However, some delivery quality indicators across the various dimensions are negatively correlated with delivery quality. In other words, a higher value indicates worse delivery quality, and a lower value indicates better delivery quality. Examples include average delivery time per order, 55-minute delay rate, and ETA+15 delay rate. If a delivery quality indicator in a certain dimension is negatively correlated with delivery quality, it is processed before weighted summation based on influence weights. This processed indicator is then positively correlated with delivery quality. Finally, a weighted summation is performed based on the influence weights and the processed indicator. For example, taking the negative value of the delivery quality indicator in a negatively correlated dimension can make the processed indicator positively correlated with delivery quality.

[0139] In some embodiments, before determining delivery quality parameters based on delivery quality indicators in multiple dimensions, the delivery quality indicators in multiple dimensions are normalized or standardized to unify the units or magnitudes of the delivery quality indicators in multiple dimensions with different units or magnitudes, so that the processed delivery quality indicators in multiple dimensions can be weighted and summed.

[0140] In another possible implementation, the delivery quality parameter is one of multiple dimensions of delivery quality indicators. For example, based on the needs of the actual application scenario, any one of the multiple dimensions of delivery quality indicators can be selected as the delivery quality parameter; this application embodiment does not impose any restrictions on this.

[0141] 303. Based on multiple first-weighted reassemblies, the delivery quality parameters corresponding to the multiple first-weighted reassemblies, and the timeliness ratios corresponding to the multiple first-weighted reassemblies, the server obtains the first distribution of the delivery quality parameters corresponding to the multiple second-weighted reassemblies and the second distribution of the timeliness ratios corresponding to the multiple second-weighted reassemblies.

[0142] It should be noted that one first weighting reassembly corresponds to one delivery quality parameter, and one first weighting reassembly corresponds to one timeliness ratio. One second weighting reassembly corresponds to one delivery quality parameter, and one second weighting reassembly corresponds to one timeliness ratio. The first distribution includes multiple delivery quality parameters corresponding to each second weighting reassembly and the probability of achieving each delivery quality parameter given the application of each second weighting reassembly. That is, the first distribution includes multiple second weighting reassemblies, multiple delivery quality parameters, and multiple probabilities. Specifically, one second weighting reassembly corresponds to multiple delivery quality parameters, and each delivery quality parameter corresponding to one second weighting reassembly corresponds to one probability, which represents the likelihood of achieving that delivery quality parameter given the application of that second weighting reassembly.

[0143] Similarly, the second distribution includes multiple time efficiency ratios corresponding to each second weighting reorganization and the probability of achieving each time efficiency ratio given the application of each second weighting reorganization. That is, the second distribution includes multiple second weighting reorganizations, multiple time efficiency ratios, and multiple probabilities. Specifically, one second weighting reorganization corresponds to multiple time efficiency ratios, and one time efficiency ratio corresponding to one second weighting reorganization corresponds to one probability, which represents the likelihood of achieving that time efficiency ratio given the application of that second weighting reorganization.

[0144] Furthermore, in on-demand delivery scenarios, the applied weighted reassemblies affect the overall matching degree between waybills and transportation capacity. This overall matching degree significantly impacts delivery quality and timeliness. If a large number of weighted reassemblies are tested in a production environment to collect experimental data, and the target weighted reassemblies are determined based on this data, it would pose a significant risk to actual delivery operations, exacerbating operational instability. Therefore, to reduce the risks to actual delivery operations and maintain operational stability, the first and second distributions corresponding to a large number of second weighted reassemblies are predicted based on the delivery quality parameters and timeliness of a small number of first weighted reassemblies. The target weighted reassemblies are then determined based on these first and second distributions. In other words, the first and second distributions corresponding to a large number of second weighted reassemblies are predicted based on the delivery quality parameters and timeliness of a small number of first weighted reassemblies, and the number of multiple first weighted reassemblies is less than the number of multiple second weighted reassemblies.

[0145] Furthermore, in on-demand delivery scenarios, there are many uncertainties such as order rejections and variable attendance times. It is difficult for simulation systems to accurately simulate delivery quality parameters and timeliness ratios under various weightings. Therefore, by abstractly modeling the problem of maintaining the stability of the target capacity type while ensuring delivery quality, predictions are made based on the established mathematical model and the known delivery quality parameters and timeliness ratios corresponding to multiple first weightings to obtain the first and second distribution scenarios.

[0146] The mathematical model established based on the problem of maintaining the stability of the target capacity type while ensuring delivery quality is shown in the following formula 2:

[0147] Formula 2:

[0148] Where, obj represents the delivery quality parameter; n represents the number of delivery quality indicators across multiple dimensions, where n is an integer greater than or equal to 1; w i f represents the influence weight of the delivery quality indicator in the i-th dimension; i is an integer greater than or equal to 1 and less than or equal to n; i (α,β) represents the delivery quality index achieved in the i-th dimension when the first weighting is applied; α is the first weight in the first weighting; β is the second weight in the first weighting, and α and β are decision variables. st, or subject to, represents the constraint condition; c lp (α,β) represents the timeliness of the target capacity type; c ks (α,β) represents the timeliness of the capacity referenced by the capacity type; c lp (α, β) / c ks (α,β) represents the time efficiency ratio; k represents the time efficiency ratio threshold, k is a value greater than 1, and the specific value of k can be flexibly configured, for example, k can be 1.2, 1.5 or 2, etc., and this application embodiment does not limit this.

[0149] In real-world environments, the delivery quality parameters and timeliness ratios corresponding to each weighted reorganization exhibit randomness. The first distribution of delivery quality parameters and the second distribution of timeliness ratios corresponding to multiple weighted reorganizations both conform to a Gaussian distribution. These distributions can be viewed as joint Gaussian distributions of the weights within the reorganization, as shown in the heatmap of this joint Gaussian distribution. Figure 4As shown. Since both the first distribution of delivery quality parameters and the second distribution of timeliness ratio can be considered as a joint Gaussian distribution of multiple weights in the weighted reassembly, Gaussian process regression can be used to fit these distributions. That is, step 303 includes: the server performing Gaussian process regression based on multiple first weighted reassemblies and the corresponding delivery quality parameters to obtain the first distribution of delivery quality parameters corresponding to multiple second weighted reassemblies; and the server performing Gaussian process regression based on multiple first weighted reassemblies and the corresponding timeliness ratios to obtain the second distribution of timeliness ratios corresponding to multiple second weighted reassemblies.

[0150] The above technical solution, through Gaussian process regression, fits the first distribution of delivery quality parameters and the second distribution of timeliness ratio corresponding to a large number of second weighted reassemblies based on a small number of first weighted reassemblies. This eliminates the need to test a large number of weighted reassemblies in the production environment, reducing the risk to actual delivery operations caused by determining the target weighted reassemblies and improving the stability of delivery operations. Furthermore, compared to simulating delivery quality parameters and timeliness ratios under various weighted reassemblies through a simulation system, this solution obtains the distribution of delivery quality parameters and timeliness ratios based on the objectively consistent probability distribution, thus improving the accuracy of obtaining the distribution of delivery quality parameters and timeliness ratios.

[0151] The first distribution of delivery quality parameters corresponding to multiple second weightings is a multidimensional Gaussian distribution. The distribution of delivery quality parameters corresponding to the first weight in the weighting reorganization can be considered as a one-dimensional Gaussian distribution within this multidimensional Gaussian distribution, and the distribution of delivery quality parameters corresponding to the second weight in the weighting reorganization can be considered as a Gaussian distribution with another dimension within this multidimensional Gaussian distribution. Similarly, the second distribution of timeliness ratios corresponding to multiple second weightings is a multidimensional Gaussian distribution. The distribution of timeliness ratios corresponding to the first weight in the weighting reorganization can be considered as a one-dimensional Gaussian distribution within this multidimensional Gaussian distribution, and the distribution of timeliness ratios corresponding to the second weight in the weighting reorganization can be considered as a Gaussian distribution with another dimension within this multidimensional Gaussian distribution.

[0152] In one example, the distribution of the delivery quality parameter corresponding to the first weight is as follows: Figure 5As shown, the horizontal axis represents the value of the first weight, and the vertical axis represents the value of the delivery quality parameter. A point in the shaded area corresponds to a value of the first weight and a value of the delivery quality parameter, representing the delivery quality parameter achieved when applying the first weight. Additionally, each point corresponds to a probability, representing the probability of achieving the delivery quality parameter when applying the first weight. The curve passing through the shaded area is the mean curve. A point on this mean curve corresponds to a value of the first weight and a value of the delivery quality parameter. The delivery quality parameter value corresponding to a point on the mean curve is the average of multiple delivery quality parameters corresponding to the first weight at that point. "+" represents the initial point, which is determined based on the known first weight recombination and the delivery quality parameter corresponding to the first weight recombination. Points in the shaded area other than those with "+" represent points determined based on the second weight recombination and the predicted delivery quality parameter corresponding to the second weight recombination. Figure 5 In the distribution shown, points near the initial point have higher certainty, while points farther from the initial point have higher uncertainty. Similarly, the distributions of the timeliness ratio corresponding to the first weight, the distributions of the delivery quality parameters corresponding to the second weight, and the distributions of the timeliness ratio corresponding to the second weight are all consistent with... Figure 5 The distributions shown are similar, and will not be illustrated with examples here.

[0153] 304. Based on the first distribution and the second distribution of multiple second weighted reorganizations, the server determines the target weighted reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from the multiple second weighted reorganizations.

[0154] In some embodiments, the server uses the largest delivery quality parameter among the delivery quality parameters corresponding to multiple first weighted reorganizations as a reference value to determine whether the delivery quality parameter has improved. This is done by referring to the largest delivery quality parameter among the multiple first weighted reorganizations to constrain the improvement in delivery quality of the determined target weighted reorganization. Step 304 above includes steps 3041 to 3042:

[0155] 3041. Based on the first distribution of multiple second weight reorganizations, the second distribution of multiple second weight reorganizations, and reference delivery quality parameters, the server determines the reference values ​​for weight selection of multiple second weight reorganizations.

[0156] Among them, the reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to multiple first weighted reorganizations, and the reference value for weight selection is a comprehensive index that integrates the timeliness ratio constraint and the delivery quality parameter.

[0157] In some embodiments, Bayesian optimization is employed to determine the target weight reassembly that satisfies the timeliness ratio constraint and maximizes the corresponding delivery quality parameter, given its suitability for solving complex optimization problems with unknown objective function expressions, non-convexity, multimodality, and high evaluation costs. In Bayesian optimization, a reference value for weight selection is determined based on the acquisition function, and then the target weight reassembly that maximizes this reference value is selected as the optimal solution under the current distribution.

[0158] Optionally, the data acquisition function, centered on the concept of EI (Expected Improvement), evaluates potential optimal points and selects the point that brings the greatest expected improvement as the optimal point. In this embodiment, to ensure delivery quality while guaranteeing that the delivery time of the target capacity type is greater than that of the reference capacity type, a data acquisition function is used that includes terms for evaluating expected improvement and terms for constraining the delivery time ratio. This function determines the weighted reference values ​​that integrate both the delivery time ratio constraint and the delivery quality parameters.

[0159] In this embodiment, the step of determining the weight selection reference value using a data acquisition function that includes an evaluation term for expected improvement and a term for constraining the timeliness ratio includes: for any second weight reassembly among multiple second weight reassemblies, the server determines the expected delivery quality corresponding to the second weight reassembly based on the first distribution of the multiple second weight reassemblies and reference delivery quality parameters; the server determines the probability of condition satisfaction corresponding to the second weight reassembly based on the second distribution of the multiple second weight reassemblies; the server determines the weight selection reference value corresponding to the second weight reassembly by multiplying the expected delivery quality by the probability of condition satisfaction. The data acquisition function is shown in Formula 3 below, where α EI (x|f * c P(c(x)≥k) is the term used to evaluate the expected improvement, and P(c(x)≥k) is the term used to constrain the timeliness ratio. The server determines the reference value for selecting the weight corresponding to the second weight reorganization using the following formula three:

[0160] Formula 3: α EIC (x|f * c )=α EI (x|f * c )×P(c(x)≥k)

[0161] Where, α EIC (x|f * c ) represents the reference value for weight selection; x represents any second weight reorganization; f * c Indicates reference delivery quality parameters; α EIC (x|f* c This means that, given that the reference delivery quality parameters are already determined, the weights corresponding to any second weighting reassembly are selected from reference values; α EI (x|f * c ) represents the expected delivery quality, which is the expected delivery quality corresponding to any second weighting under the condition that the reference delivery quality parameters have been determined; c(x) represents the timeliness ratio corresponding to the second weighting; k represents the timeliness ratio threshold; P(c(x)≥k) represents the probability of the condition being met, which is the probability that the timeliness ratio achieved under the premise of applying the second weighting is greater than the timeliness ratio threshold.

[0162] In some embodiments, to increase the influence of the timeliness ratio constraint on the reference value for weight selection, α is also adjusted. EI (x|f * c The conditions α and P(c(x)≥k) are processed to make the probability of satisfying the timeliness ratio constraint greater than the order of magnitude of the expected delivery quality. For example, in α EI (x|f * c Without changing the condition, multiply P(c(x)≥k) by 10.

[0163] The aforementioned expected delivery quality represents the improvement of the delivery quality parameter achieved under the second weighting reorganization compared to the reference delivery quality parameter. The server determines the expected delivery quality corresponding to the second weighting reorganization based on the first distribution of multiple second weighting reorganizations, including: The server, based on the first distribution of multiple second weighting reorganizations, obtains multiple target delivery quality parameters that are greater than the reference delivery quality parameter from the multiple delivery quality parameters corresponding to the second weighting reorganization, along with the probabilities corresponding to the multiple target delivery quality parameters. The probability corresponding to any target delivery quality parameter refers to the probability of achieving that target delivery quality parameter under the second weighting reorganization; for any target delivery quality parameter among the multiple target delivery quality parameters, the difference between the target delivery quality parameter and the reference delivery quality parameter is determined; the product of the difference corresponding to the target delivery quality parameter and the probability corresponding to the target delivery quality parameter is determined; the mean of the products corresponding to the multiple target delivery quality parameters is determined as the expected configuration quality corresponding to the second weighting reorganization. That is, the server determines the expected delivery quality corresponding to the second weighting reorganization using the following formula:

[0164] Formula 4: α EI (x|f * c )=E[max(0,yf * c )|y~g(x|D f )]

[0165] Where, α EI (x|f * c ) represents the expected delivery quality; x represents the second weighting; f * c This represents the reference delivery quality parameter; y refers to any one of the multiple delivery quality parameters corresponding to the second weighted reassembly; max(0, yf) * c ) represents taking 0 and yf * c The maximum value in, that is, if y is greater than f * c Then max(0, yf) * c Take yf * c If y is less than or equal to f * c Then max(0,yf) * c D is 0; f This represents the set of points corresponding to multiple first-weighted reassemblies. The x-coordinate of a point corresponding to a first-weighted reassembly is the value of that first-weighted reassembly, and the y-coordinate is the delivery quality parameter corresponding to that first-weighted reassembly; g(x|D f ) represents the first distribution of delivery quality parameters corresponding to the second weighting, given that the points corresponding to multiple first weighting reorganizations are already determined; y~g(x|D f ) indicates that the delivery quality parameter corresponding to the second weighted reorganization x follows the pattern g(x|D f The distribution is represented by ); max(0, yf * c )|y~g(x|D f ) represents taking 0 and yf * c The maximum value in the middle is then compared with g(x|D) f The probability of (x, y) in the distribution is multiplied, where (x, y) is the probability of achieving the delivery quality parameter y when the second weighting of x is applied. E[] represents taking the expectation, that is, determining the max(0, yf) corresponding to multiple delivery quality parameters. * c )|y~g(x|D f The mean of the values ​​of ).

[0166] The probability of satisfying the above conditions refers to the probability that the timeliness ratio achieved under the premise of applying the second weighting reorganization meets the timeliness ratio constraint. The steps of the server determining the probability of satisfying the conditions corresponding to the second weighting reorganization based on the second distribution of multiple second weighting reorganizations include: the server, based on the second distribution of multiple second weighting reorganizations, determines multiple target timeliness ratios that are greater than the timeliness ratio threshold from the multiple timeliness ratios corresponding to the second weighting reorganizations; the sum of the probabilities corresponding to the multiple target timeliness ratios is determined as the probability of satisfying the conditions; wherein, the probability corresponding to any target timeliness ratio refers to the probability of achieving that target timeliness ratio under the premise of applying the second weighting reorganization.

[0167] 3042. The server determines the target weight reorganization with the largest reference value from multiple second-weight reorganizations.

[0168] The server determines the largest weight selection reference value from multiple second-weighted reorganizations and identifies the second-weighted reorganization corresponding to the largest weight selection reference value as the target weighted reorganization.

[0169] In one example Figure 5 The point represented by “○” is the optimal point determined through the above steps. The horizontal axis of the optimal point is the value of the target weight group, and the vertical axis is the maximum delivery quality parameter corresponding to the target weight group.

[0170] It should be noted that if the reference value for the largest weight corresponding to multiple second weight reassemblies is less than or equal to 0, it means that the difference between the delivery quality parameter corresponding to multiple second weight reassemblies and the reference delivery quality parameter is negative, that is, the delivery quality parameter corresponding to multiple second weight reassemblies is less than the reference delivery quality parameter, or the probability that the timeliness ratio corresponding to multiple second weight reassemblies is greater than the timeliness ratio threshold is 0, that is, the timeliness ratio corresponding to multiple second weight reassemblies does not meet the timeliness ratio constraint. In this case, the first weight reassembly corresponding to the reference delivery quality parameter is the current optimal weight reassembly that satisfies both the condition of having the largest delivery quality parameter and the timeliness ratio constraint. Therefore, when the reference value for the largest weight corresponding to multiple second weight reassemblies is less than or equal to 0, the first weight reassembly corresponding to the reference delivery quality parameter is determined as the target weight reassembly, and the updating and adjustment of the target weight group is suspended.

[0171] 305. When the server receives an allocation request for any waybill, it determines the allocation parameters of the waybill for multiple transport capacities based on the target weighting reorganization. The allocation parameters of any transport capacity are used to represent the overall matching degree between the waybill and the transport capacity.

[0172] After determining the target weighting reorganization, the current weighting reorganization used for capacity allocation is replaced with the target weighting reorganization, and capacity allocation is performed using the target weighting reorganization. Specifically, an allocation request for any waybill requests that the waybill be assigned to a specific capacity, enabling that capacity to deliver goods based on the waybill. If a waybill can be assigned to multiple capacities, the server determines the allocation parameters for that waybill across these multiple capacities based on the target weighting reorganization, and then determines the most suitable capacity to process the waybill based on these allocation parameters.

[0173] In some embodiments, the target weighting reorganization includes a first weight and a second weight. The first weight represents the importance of the delivery matching degree between capacity and waybill to capacity allocation, and is used to weight the delivery matching parameter representing the delivery matching degree between capacity and waybill. The second weight represents the importance of capacity type to capacity allocation, and is used to weight the level parameter representing the capacity type. Based on this, the server determines the allocation parameters of the waybill for multiple capacity based on the target weighting reorganization, including: for any capacity among the multiple capacity options, the server obtains the delivery matching parameter of the capacity, and obtains the level parameter corresponding to the capacity type to which the capacity belongs, wherein the delivery matching parameter represents the delivery matching degree between the capacity and waybill; the server determines a first product of the delivery matching parameter of the capacity and the first weight in the target weighting reorganization; determines a second product of the level parameter of the capacity and the second weight in the target weighting reorganization; and determines the sum of the first product and the second product as the allocation parameter of the capacity. That is, the server determines the allocation parameters between any capacity and waybill using the following formula:

[0174] Formula 5: Cost(Rider) i ,waybill j )=α×Score ij +β×Level i

[0175] Among them, Rider i Indicates transport capacity i; waybill j Represents waybill j; Cost(Rider) i ,waybill j ) represents the allocation parameter, used to indicate the overall matching degree between capacity i and waybill j; α represents the first weight; Score ij Level represents the degree of delivery matching between capacity i and waybill j; β represents the second weight; i This represents the level parameter of transport capacity i.

[0176] The aforementioned delivery matching parameters are used to represent the degree of delivery matching between transportation capacity and waybills. In some embodiments, the delivery matching parameters are determined based on at least one of a timeout index, a route index, an order acceptance index, and a order volume index. The timeout index represents the probability that the waybill will be delivered late if handled by the transportation capacity. The route index represents the degree of overlap between the current route of the transportation capacity and the route required to complete the waybill. The order acceptance index represents the proportion of waybills allocated to the transportation capacity that are accepted and completed by the transportation capacity. The order volume index represents the number of waybills currently being delivered by the transportation capacity. The process of determining the delivery matching parameters based on at least one of the timeout index, route index, order acceptance index, and order volume index is similar to the process of determining delivery quality parameters based on multiple dimensions of delivery quality indicators, and will not be elaborated further here.

[0177] The grade parameters corresponding to the aforementioned capacity types are used to indicate the degree of importance attached to that capacity type during the capacity allocation process. These grade parameters affect the degree to which waybills are tilted towards that capacity type. In some embodiments, the allocation parameter is positively correlated with the overall matching degree between capacity and waybill; that is, the larger the allocation parameter, the higher the overall matching degree between capacity and waybill, and the greater the likelihood that the capacity will be identified as the target capacity. In other embodiments, the allocation parameter is negatively correlated with the overall matching degree between capacity and waybill; the smaller the allocation parameter, the higher the overall matching degree between capacity and waybill, and the greater the likelihood that the capacity will be identified as the target capacity.

[0178] If the allocation parameters are positively correlated with the overall matching degree between capacity and waybills, then the grade parameter corresponding to the capacity type is positively correlated with the degree to which waybills are allocated to that capacity type. For example, the degree to which waybills are allocated to the target capacity type should be greater than the degree to which they are allocated to the reference capacity type. Therefore, the grade parameter corresponding to the target capacity type should be greater than the grade parameter corresponding to the reference capacity type, making the grade parameter of the target capacity type larger. This, in turn, makes the allocation parameter corresponding to the target capacity type larger, causing the allocation of waybills to be tilted towards the target capacity type. If the allocation parameters are negatively correlated with the overall matching degree between capacity and waybills, then the grade parameter corresponding to the capacity type is negatively correlated with the degree to which waybills are allocated to that capacity type. For example, the grade parameter corresponding to the target capacity type should be smaller than the grade parameter corresponding to the reference capacity type.

[0179] 306. The server determines the target capacity whose allocation parameters meet the allocation conditions from multiple capacity options.

[0180] The allocation criteria are used to indicate the highest overall match between capacity and waybill. In some embodiments, the allocation parameter is positively correlated with the overall match between capacity and waybill. To determine the capacity with the highest overall match with the waybill, the allocation criterion should be set to the maximum allocation parameter. Accordingly, the server determines the target capacity with the maximum allocation parameter from multiple capacity options.

[0181] In some embodiments, the allocation parameter is negatively correlated with the overall matching degree between capacity and waybill; therefore, the allocation condition is the minimum allocation parameter. Accordingly, the server determines the target capacity with the minimum allocation parameter from multiple capacity options. In some embodiments, the server determines the target capacity with the minimum allocation parameter from multiple capacity options based on the following formula:

[0182] Formula Six:

[0183] Where w represents a waybill; r represents capacity; W represents the set of waybills to be allocated; R represents the set of capacity; x w,r Indicates whether the capacity has been determined as the target capacity, x w,r ={0,1} means x w,r The value of x is either 0 or 1. If the capacity is determined to be the target capacity, then x w，r The value is 1; if it is uncertain whether the capacity is the target capacity, then x is 1. w,r =0; C w,r Indicates the allocation parameter; min∑ w∈W ∑ r∈R x w,r ×C w,r This indicates that the target capacity is the capacity that minimizes the allocation parameters; st∑ r∈R x w,r =1, This indicates that the constraint condition is that for any waybill, the amount of capacity allocated to that waybill is 1.

[0184] The technical solution provided in this application sets a timeliness ratio constraint to ensure that the timeliness of the target capacity type is greater than that of the reference capacity type, thus maintaining the bias of waybills towards the target capacity type. Simultaneously, it sets a condition for maximizing the delivery quality parameter to guarantee delivery quality. Based on this, referencing the delivery quality parameters and timeliness ratios achieved when applying the first weighting reorganization in historical time periods, a target weighting reorganization that satisfies both the timeliness ratio constraint and the maximum delivery quality parameter is determined from multiple unapplied second weighting reorganizations. This target weighting reorganization is then applied for capacity allocation. This capacity allocation process incorporates considerations of both delivery quality and the bias towards the target capacity type, ensuring delivery quality while maintaining the bias of waybill allocation towards the target capacity type. This reduces the loss of target capacity type capacity and improves the stability of the target capacity type.

[0185] 307. If the duration for which the target weight reorganization is applied to capacity allocation reaches the target duration and the feedback adjustment conditions are met, the server will treat the target weight reorganization as the first weight reorganization and adjust the target weight reorganization based on the feedback from the first weight group.

[0186] The target duration can be flexibly configured, for example, the target duration is 1 day, 3 days or 1 week, etc., and this application embodiment does not limit it.

[0187] After the target weight reorganization has been applied to capacity allocation for the target duration, the server, before using it as the first weight reorganization, determines whether the feedback adjustment conditions are met. If the conditions are met, the target weight reorganization is used as the first weight reorganization, and adjustments are made based on the feedback from the first weight group. If the conditions are not met, adjustments to the target weight group are paused, and capacity allocation continues to be applied using the target weight reorganization.

[0188] The feedback adjustment conditions include at least one of a first feedback adjustment condition and a second feedback adjustment condition. The first feedback adjustment condition means that the number of feedback adjustments has not reached the upper limit. The second feedback adjustment condition means that the delivery quality parameter achieved when applying the target weight reorganization is greater than the reference delivery quality parameter. The upper limit for the number of feedback adjustments can be flexibly configured as any positive integer, such as 3, 5, or 6, and this embodiment does not impose such a limitation. The delivery quality parameter achieved when applying the target weight reorganization refers to the actual delivery quality parameter achieved when applying the target weight reorganization. The reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to multiple first weight reorganizations.

[0189] In some embodiments, the feedback adjustment conditions include a first feedback adjustment condition. If the number of feedback adjustments has not reached the upper limit, the server will take the target weight reorganization as the first weight reorganization and adjust the target weight reorganization based on the feedback of the first weight group. If the number of feedback adjustments has reached the upper limit, the adjustment of the target weight group will be suspended.

[0190] In some embodiments, the feedback adjustment conditions include a second feedback adjustment condition. If the delivery quality parameter reached when applying the target weight reorganization is greater than the reference delivery quality parameter, the server will treat the target weight reorganization as the first weight reorganization and adjust the target weight reorganization based on the feedback from the first weight group. If the delivery quality parameter reached when applying the target weight reorganization is less than or equal to the reference delivery quality parameter, the adjustment of the target weight group will be suspended.

[0191] In some embodiments, the feedback adjustment conditions include a first feedback adjustment condition and a second feedback adjustment condition. If the first feedback adjustment condition and the second feedback adjustment condition are met, the server will take the target weight reorganization as the first weight reorganization and adjust the target weight reorganization based on the feedback from the first weight group. If either the first feedback adjustment condition or the second feedback adjustment condition is not met, the adjustment of the target weight group will be suspended.

[0192] In this process, the server adjusts the target weight reorganization based on the feedback from the first weight group, following a similar procedure to steps 301 to 306 above. That is, after treating the target weight reorganization as the first weight reorganization, the server continues execution from step 301. See one example. Figure 6 , Figure 6 The point represented by "○" is the optimal point determined through feedback adjustment. The horizontal axis of this optimal point is the value of the adjusted target weight group, and the vertical axis is the maximum delivery quality parameter corresponding to the adjusted target weight group.

[0193] In some embodiments, after the target weight restructuring is suspended for a reference period, the target weight restructuring is adjusted again through a process similar to steps 301 to 306 above, so that the applied target weight restructuring is continuously updated as the actual application scenario changes, ensuring that the target weight restructuring adapts to the changes in the current application scenario.

[0194] It should be noted that the process of determining the target weight reorganization as the first weight reorganization and adjusting the target weight reorganization based on the first weight reorganization is also part of the Bayesian optimization method. By applying the Bayesian optimization method, the optimal target weight reorganization can be obtained after fewer feedback adjustments, thereby reducing the number of times non-optimal weight reorganizations are used in the actual application environment, reducing the impact of the determination of the target weight group on the actual application environment, reducing the risks to the actual delivery business, and improving the stability of the delivery business operation.

[0195] The above technical solution, combining Gaussian process regression, Bayesian optimization, and feedback adjustment methods, achieves adaptive adjustment of weights under hybrid capacity scheduling. Under the constraint of ensuring the timeliness of the capacity, it maximizes the optimization of the delivery experience. Compared with the solution of manually adjusting weights based on personal experience, it is more accurate and has a lower probability of failing to meet the timeliness ratio constraint or the condition of maximizing the delivery quality parameter after weight adjustment. It can better meet the need to maintain the stability of the target capacity type while ensuring delivery quality.

[0196] To facilitate understanding, the following will be combined with... Figure 7 This section explains the process of determining, applying, and adjusting the target weight group. (See also...) Figure 7The determination and feedback adjustment of the target weight group are performed offline. The online capacity allocation process is not affected during the determination of the target weight reorganization and the feedback adjustment of the target weight reorganization. In this embodiment, it is assumed that the prior distribution of the target and constraints, i.e., the delivery quality parameters and timeliness ratio, conform to a Gaussian distribution. Before the first determination of the target weight reorganization, multiple initial first weight reorganizations need to be selected. After selecting the initial multiple first weight reorganizations, it is determined whether the feedback adjustment conditions are met. If the feedback adjustment conditions are met, the posterior probability distribution of the target and constraints is updated using existing data through step 303 above. That is, based on the multiple first weight reorganizations, the delivery quality parameters corresponding to the multiple first weight reorganizations, and the timeliness ratios corresponding to the multiple first weight reorganizations, the delivery quality parameters corresponding to the multiple second weight reorganizations are obtained. The first distribution and the second distribution of the timeliness ratio corresponding to multiple second weighting groups; after updating the posterior probability distribution of the target and constraints with existing data, through step 304 above, the next optimal weighting group is selected using the collection function, that is, the target weighting group that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter is determined; after determining the target weighting group, the target weighting group is applied to the online capacity allocation process, that is, through step 305, the allocation parameters of multiple capacities are determined, and then through step 306, the target capacity whose allocation parameters satisfy the allocation conditions is determined, that is, the waybill and capacity are matched. In the process of applying the target weighting group, through steps 301 to 302, the delivery quality parameters and timeliness ratio generated when applying the target weighting group can be obtained; then through step 307, when the time for which the target weighting group is applied to the capacity allocation reaches the target time and the feedback adjustment condition is met, the target weighting group is used as the first weighting group, and the above process is repeated. Based on the feedback of the first weighting group, the target weighting group is adjusted, and the adjusted target weighting group is used for capacity allocation.

[0197] In some embodiments, before determining the allocation parameters based on information about capacity and waybills, multiple currently online capacity units are filtered. This initial screening removes capacity units with low matching degrees to the waybill delivery, reducing the number of capacity units requiring allocation parameter determination, lowering computational resource consumption, and improving capacity allocation efficiency. In on-demand delivery scenarios, multiple capacity units are filtered based on at least one of the following: delivery distance, range of the capacity's permanent locations, degree of overlap between the capacity's current route and the route required to complete the waybill, and probability of waybill delivery timeout.

[0198] In some embodiments, in order to cover more second weightings by the distribution determined based on the initial multiple first weighting reorganizations, multiple weighting reorganizations with large spans are selected as the initial multiple first weighting reorganizations. That is, multiple first weighting reorganizations with large differences are selected as the initial multiple first weighting reorganizations. For example, in one first weighting reorganization, the first weight is 0 and the second weight is 1; in another weighting reorganization, the first weight is 0.5 and the second weight is 0.5; and in yet another first weighting reorganization, the first weight is 1 and the second weight is 0.

[0199] In some embodiments, the method for determining the target capacity described above can be applied according to regional divisions. For example, the method for determining the target capacity described above can be applied within a regional area of ​​a city, where the multiple capacities are capacities within that city, and the waybills are waybills within that city.

[0200] It should be noted that the above method for determining target capacity has been applied in multiple cities. In practical application, the timeliness ratio has remained relatively stable, the order completion rate has increased by 0.04–0.29 pp (percent points), the relative on-time rate has increased by 0.16–0.31 pp, the 55-minute delay rate has decreased by 0.36–1.06 pp, the average delivery time per order has decreased by 0.2–0.79 minutes, and the ETA+15 delay rate has decreased by 0.09–0.18 pp. A comparison chart of the relative on-time rate before and after applying the above method for determining target capacity in one city is shown below. Figure 8 As shown in the figure, a comparison chart of the percentage of daily delivery orders exceeding 55 minutes before and after applying the above method for determining target capacity in a city is shown. Figure 9 As shown in the figure, a comparison chart of the average delivery time per order before and after applying the above method for determining target transportation capacity in a city is shown. Figure 10 As shown in the figure, by applying the above method for determining the target capacity, while maintaining the capacity tilt of waybills towards the target capacity type, the delivery quality is improved, thereby improving the delivery quality while maintaining the stability of the target capacity type.

[0201] All of the above-mentioned optional technical solutions can be combined in any way to form the optional embodiments of this application, and will not be described in detail here.

[0202] Figure 11 This is a block diagram of a device for determining target transport capacity provided in an embodiment of this application. See also... Figure 11 The device includes:

[0203] The timeliness ratio determination module 1101 is used to determine the timeliness ratio corresponding to the first weight reorganization as the ratio of the timeliness of the target capacity type when the first weight reorganization is applied to the timeliness of the reference capacity type when the first weight reorganization is applied for any first weight reorganization in multiple first weight reorganizations applied in history to capacity allocation.

[0204] The quality parameter acquisition module 1102 is used to acquire the delivery quality parameters achieved when the first weighting is applied;

[0205] The target weight reorganization determination module 1103 is used to determine the target weight reorganization that meets the timeliness ratio constraint and has the largest corresponding delivery quality parameter from multiple second weight reorganizations that have not yet been applied to capacity allocation, based on the timeliness ratio and delivery quality parameters corresponding to multiple first weight reorganizations. The timeliness ratio constraint is that the timeliness ratio is greater than or equal to the timeliness ratio threshold.

[0206] The allocation parameter determination module 1104 is used to determine the allocation parameters of the waybill for multiple transport capacities based on the target weight reorganization when receiving an allocation request for any waybill. The allocation parameters of any transport capacity are used to represent the overall matching degree between the waybill and the transport capacity.

[0207] The target capacity determination module 1105 is used to determine the target capacity whose allocation parameters meet the allocation conditions from multiple capacity options;

[0208] Among them, timeliness is used to represent the number of waybills delivered per unit of capacity per unit of time; any weighting group includes a first weight and a second weight applied when allocating capacity. The first weight is used to represent the importance of the degree of matching between capacity and waybill in capacity allocation, and the second weight is used to represent the importance of capacity type in capacity allocation.

[0209] The apparatus provided in this application sets a timeliness ratio constraint to ensure that the timeliness of the target capacity type is greater than that of the reference capacity type, thereby maintaining the bias of waybills towards the target capacity type. Simultaneously, it sets a condition to maximize the delivery quality parameter to guarantee delivery quality. Based on this, referencing the delivery quality parameters and timeliness ratios achieved when applying the first weighting reorganization in historical time periods, it determines a target weighting reorganization from multiple unapplied second weighting reorganizations that satisfies both the timeliness ratio constraint and the maximum delivery quality parameter. This target weighting reorganization is then applied for capacity allocation. This capacity allocation process incorporates considerations of both delivery quality and the bias towards the target capacity type, ensuring delivery quality while maintaining the bias of waybill allocation towards the target capacity type. This reduces the loss of target capacity type capacity and improves the stability of the target capacity type.

[0210] In one possible implementation, the target weight reorganization determination module 1103 includes:

[0211] The distribution acquisition unit is used to acquire, based on multiple first weighted reorganizations, delivery quality parameters corresponding to multiple first weighted reorganizations, and timeliness ratios corresponding to multiple first weighted reorganizations, the first distribution of delivery quality parameters corresponding to multiple second weighted reorganizations and the second distribution of timeliness ratios corresponding to multiple second weighted reorganizations;

[0212] The target weight reorganization determination unit is used to determine the target weight reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from multiple second weight reorganizations based on the first distribution situation and the second distribution situation corresponding to multiple second weight reorganizations.

[0213] The first distribution includes multiple delivery quality parameters corresponding to each second weighting and the probability of achieving each delivery quality parameter under the premise of applying each second weighting.

[0214] The second distribution includes multiple time-efficiency ratios corresponding to each second weighting reorganization and the probability of achieving each time-efficiency ratio under the premise of applying each second weighting reorganization.

[0215] In another possible implementation, the distribution acquisition unit is used for:

[0216] Based on multiple first-weighted reorganizations and the delivery quality parameters corresponding to multiple first-weighted reorganizations, Gaussian process regression is performed to obtain the first distribution of delivery quality parameters corresponding to multiple second-weighted reorganizations.

[0217] Based on multiple first-weighted reorganizations and their corresponding time efficiency ratios, Gaussian process regression is performed to obtain the second distribution of the time efficiency ratios corresponding to multiple second-weighted reorganizations.

[0218] In another possible implementation, the target weight reorganization determination unit includes:

[0219] The reference value determination subunit is used to determine the weight selection reference value for multiple second weight reorganizations based on the first distribution corresponding to multiple second weight reorganizations, the second distribution corresponding to multiple second weight reorganizations, and the reference delivery quality parameters.

[0220] The target weight reorganization determination sub-unit is used to determine the target weight reorganization with the largest reference value among multiple second weight reorganizations;

[0221] Among them, the reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to multiple first weighted reorganizations, and the reference value for weight selection is a comprehensive index that integrates the timeliness ratio constraint and the delivery quality parameter.

[0222] In another possible implementation, the reference value determines the sub-unit for:

[0223] For any second weighting reorganization among multiple second weighting reorganizations, based on the first distribution corresponding to the multiple second weighting reorganizations and the reference delivery quality parameters, the expected delivery quality corresponding to the second weighting reorganization is determined. The expected delivery quality is used to represent the degree of improvement of the delivery quality parameters achieved under the premise of applying the second weighting reorganization compared to the reference delivery quality parameters.

[0224] Based on the second distribution corresponding to multiple second weight reorganizations, the probability of condition satisfaction corresponding to the second weight reorganization is determined. The probability of condition satisfaction refers to the probability that the time efficiency ratio achieved under the premise of applying the second weight reorganization satisfies the time efficiency ratio constraint.

[0225] The product of the expected delivery quality and the probability of the condition being met is determined as the reference value for selecting the weights corresponding to the second weight reorganization.

[0226] In another possible implementation, the reference value determines the sub-unit for:

[0227] Based on the first distribution corresponding to multiple second weighting reorganizations, multiple target delivery quality parameters that are greater than the reference delivery quality parameters are obtained from the multiple delivery quality parameters corresponding to the second weighting reorganizations, as well as the probabilities corresponding to the multiple target delivery quality parameters. The probability corresponding to any target delivery quality parameter refers to the probability of achieving any target delivery quality parameter under the premise of applying the second weighting reorganization.

[0228] For any target delivery quality parameter among multiple target delivery quality parameters, determine the difference between the target delivery quality parameter and the reference delivery quality parameter; determine the product of the difference corresponding to the target delivery quality parameter and the probability corresponding to the target delivery quality parameter;

[0229] The mean of the products corresponding to multiple target delivery quality parameters is determined as the configuration quality expectation corresponding to the second weighted reassembly.

[0230] In another possible implementation, the reference value determines the sub-unit for:

[0231] Based on the second distribution corresponding to multiple second weighting reorganizations, multiple target timeliness ratios that are greater than the timeliness ratio threshold are determined from the multiple timeliness ratios corresponding to the second weighting reorganizations;

[0232] The sum of the probabilities corresponding to the timeliness of multiple targets is determined as the probability of condition satisfaction;

[0233] The probability corresponding to any target timeliness ratio refers to the probability of achieving any target timeliness ratio under the premise of applying the second weighted reorganization.

[0234] In another possible implementation, the device also includes:

[0235] The feedback adjustment unit is used to determine the target weight reorganization as the first weight reorganization if the duration of the target weight reorganization applied to capacity allocation reaches the target duration and the feedback adjustment conditions are met. For any first weight reorganization among multiple first weight reorganizations applied to capacity allocation in the past, the ratio of the timeliness of the capacity of the target capacity type when the first weight reorganization is applied to the timeliness of the capacity of the reference capacity type when the first weight reorganization is applied is determined as the timeliness ratio corresponding to the first weight reorganization. In addition, the unit obtains the delivery quality parameters achieved when the first weight reorganization is applied.

[0236] In another possible implementation, the feedback adjustment condition includes at least one of a first feedback adjustment condition and a second feedback adjustment condition;

[0237] The first feedback adjustment condition refers to the fact that the number of feedback adjustments has not reached the upper limit;

[0238] The second feedback adjustment condition refers to the delivery quality parameter achieved when applying the target weight reorganization being greater than the reference delivery quality parameter, where the reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to multiple first weight reorganizations.

[0239] In another possible implementation, the parameter allocation determination module 1104 is used for:

[0240] For any one of the multiple transport capacities, obtain the delivery matching parameters of the transport capacity, and obtain the level parameters corresponding to the transport capacity type to which the transport capacity belongs. The delivery matching parameters are used to indicate the degree of delivery matching between the transport capacity and the waybill.

[0241] Determine the first product of the delivery matching parameters of the transport capacity and the first weight in the target weighting reorganization;

[0242] Determine the product of the capacity level parameters and the second weight in the target weighting reorganization;

[0243] The sum of the first and second products is determined as the capacity allocation parameter.

[0244] It should be noted that the apparatus for determining target capacity provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the server can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus for determining target capacity and the method for determining target capacity provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiments, which will not be repeated here.

[0245] Figure 12This is a block diagram of a server provided in an embodiment of this application. The server 1200 can vary significantly due to different configurations or performance. It may include one or more Central Processing Units (CPUs) 1201 and one or more memories 1202. The memory 1202 stores at least one line of program code, which is loaded and executed by the processor 1201 to implement the method for determining target capacity provided in the various method embodiments described above. Of course, the server may also have wired or wireless network interfaces, a keyboard, and input / output interfaces for input and output. The server may also include other components for implementing device functions, which will not be elaborated upon here.

[0246] In an exemplary embodiment, a computer-readable storage medium is also provided, which stores at least one piece of program code that can be executed by a processor in a server to perform the method for determining target capacity in the above embodiments. For example, the computer-readable storage medium may be ROM (Read-Only Memory), RAM (Random Access Memory), CD-ROM (CompactDisc Read-Only Memory), magnetic tape, floppy disk, and optical data storage device, etc.

[0247] This application also provides a computer program product or computer program that includes computer program code stored in a computer-readable storage medium. The server's processor reads the computer program code from the computer-readable storage medium and executes the computer program code, causing the server to perform the method for determining target capacity in the above-described method embodiments.

[0248] Those skilled in the art will understand that all or part of the steps of the above embodiments can be implemented by hardware or by a program instructing related hardware. The program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk.

[0249] The above description is merely an optional embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for determining target transport capacity, characterized in that, The method includes: For any first weighting reorganization in a series of first weighting reorganizations applied to capacity allocation in the past, the ratio of the timeliness of the capacity of the target capacity type when the first weighting reorganization is applied to the timeliness of the capacity of the reference capacity type when the first weighting reorganization is applied is determined as the timeliness ratio corresponding to the first weighting reorganization, and the delivery quality parameters achieved when the first weighting reorganization is applied are obtained. Based on the multiple first weighted reorganizations, the delivery quality parameters corresponding to the multiple first weighted reorganizations, and the timeliness ratios corresponding to the multiple first weighted reorganizations, a Gaussian process regression is performed to obtain a first distribution of the delivery quality parameters corresponding to the multiple second weighted reorganizations and a second distribution of the timeliness ratios corresponding to the multiple second weighted reorganizations. The first distribution includes multiple delivery quality parameters corresponding to each second weighted reorganization and the probability of achieving each delivery quality parameter under the premise of applying each second weighted reorganization. The second distribution includes multiple timeliness ratios corresponding to each second weighted reorganization and the probability of achieving each timeliness ratio under the premise of applying each second weighted reorganization. Based on the first distribution and the second distribution of the multiple second weighted reorganizations, from the multiple second weighted reorganizations that have not yet been applied to capacity allocation, a target weighted reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter is determined. The timeliness ratio constraint is that the timeliness ratio is greater than or equal to the timeliness ratio threshold. Upon receiving an allocation request for any waybill, the allocation parameters for the waybill among multiple transport capacities are determined based on the target weighting reorganization. From the plurality of transport capacities, a target transport capacity whose allocation parameters satisfy the allocation conditions is determined, wherein the allocation parameters of any transport capacity are used to represent the overall matching degree between the waybill and the transport capacity; Among them, timeliness is used to indicate the number of waybills delivered per unit of transportation capacity per unit of time; Any weighting reorganization includes a first weight and a second weight applied during capacity allocation. The first weight is used to indicate the importance of the degree of delivery matching between capacity and waybill to capacity allocation, and the second weight is used to indicate the importance of capacity type to capacity allocation.

2. The method according to claim 1, characterized in that, The step of determining the target weighted reorganization that satisfies the timeliness ratio constraint and has the largest corresponding delivery quality parameter from among the multiple second weighted reorganizations that have not yet been applied to capacity allocation, based on the first distribution and the second distribution of the multiple second weighted reorganizations, includes: Based on the first distribution of the multiple second weighting reorganizations, the second distribution of the multiple second weighting reorganizations, and the reference delivery quality parameters, the weight selection reference values ​​for the multiple second weighting reorganizations are determined. From the plurality of second weighting reorganizations, determine the target weighting reorganization with the largest reference value for the corresponding weight selection; The reference delivery quality parameter is the largest delivery quality parameter among the multiple first weighted groups, and the reference value for weight selection is a comprehensive index that integrates the timeliness ratio constraint and the delivery quality parameter.

3. The method according to claim 2, characterized in that, The step of determining the reference values ​​for weight selection corresponding to the multiple second weighting reorganizations based on the first distribution of the multiple second weighting reorganizations, the second distribution of the multiple second weighting reorganizations, and reference delivery quality parameters includes: For any second weighting reorganization among the plurality of second weighting reorganizations, based on the first distribution corresponding to the plurality of second weighting reorganizations and the reference delivery quality parameter, the expected delivery quality corresponding to the second weighting reorganization is determined. The expected delivery quality is used to represent the degree of improvement of the delivery quality parameter achieved under the premise of applying the second weighting reorganization compared to the reference delivery quality parameter. Based on the second distribution corresponding to the multiple second weighting reorganizations, the probability of condition satisfaction corresponding to the second weighting reorganization is determined. The probability of condition satisfaction refers to the probability that the time efficiency ratio achieved under the premise of applying the second weighting reorganization satisfies the time efficiency ratio constraint condition. The product of the expected delivery quality and the probability of the condition being met is determined as the reference value for selecting the weights corresponding to the second weighted reassembly.

4. The method according to claim 3, characterized in that, The step of determining the expected delivery quality corresponding to the second weighted reassembly based on the first distribution of the plurality of second weighted reassemblies and the reference delivery quality parameters includes: Based on the first distribution corresponding to the multiple second weighting reorganizations, multiple target delivery quality parameters that are greater than the reference delivery quality parameter are obtained from the multiple delivery quality parameters corresponding to the second weighting reorganizations, and the probabilities corresponding to the multiple target delivery quality parameters are obtained. The probability corresponding to any target delivery quality parameter refers to the probability of achieving any target delivery quality parameter under the premise of applying the second weighting reorganization. For any one of the plurality of target delivery quality parameters, determine the difference between the target delivery quality parameter and the reference delivery quality parameter; Determine the product of the difference corresponding to the target delivery quality parameter and the probability corresponding to the target delivery quality parameter; The mean of the products corresponding to the multiple target delivery quality parameters is determined as the configuration quality expectation corresponding to the second weighted reassembly.

5. The method according to claim 3, characterized in that, The step of determining the probability of satisfying the condition corresponding to the second weighting based on the second distribution corresponding to the plurality of second weighting reorganizations includes: Based on the second distribution corresponding to the multiple second weighting reorganizations, multiple target time efficiency ratios that are greater than the time efficiency ratio threshold are determined from the multiple time efficiency ratios corresponding to the second weighting reorganizations; The sum of the probabilities corresponding to the multiple target timeliness ratios is determined as the probability of the condition being met; The probability corresponding to any target timeliness ratio refers to the probability of achieving any target timeliness ratio under the premise of applying the second weighting reorganization.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: If the duration for which the target weight reorganization is applied to capacity allocation reaches the target duration and the feedback adjustment conditions are met, the target weight reorganization is taken as the first weight reorganization. The following steps are performed: for any one of the multiple first weight reorganizations applied to capacity allocation in the past, the ratio of the timeliness of the capacity of the target capacity type when the first weight reorganization is applied to the timeliness of the capacity of the reference capacity type when the first weight reorganization is applied is determined as the timeliness ratio corresponding to the first weight reorganization; and the delivery quality parameters achieved when the first weight reorganization is applied are obtained.

7. The method according to claim 6, characterized in that, The feedback adjustment conditions include at least one of the first feedback adjustment conditions and the second feedback adjustment conditions; The first feedback adjustment condition refers to the fact that the number of feedback adjustments has not reached the upper limit; The second feedback adjustment condition refers to the delivery quality parameter achieved when the target weighting is applied being greater than the reference delivery quality parameter, where the reference delivery quality parameter is the largest delivery quality parameter among the delivery quality parameters corresponding to the plurality of first weightings.

8. The method according to any one of claims 1-5, characterized in that, The step of determining the allocation parameters of the waybill for multiple transport capacities based on the target weight reorganization includes: For any one of the multiple transport capacities, obtain the delivery matching parameters of the transport capacity, and obtain the level parameters corresponding to the transport capacity type to which the transport capacity belongs. The delivery matching parameters are used to represent the degree of delivery matching between the transport capacity and the waybill. Determine the first product of the delivery matching parameters of the transport capacity and the first weight in the target weighting; Determine the second product of the capacity level parameter and the second weight in the target weighting; The sum of the first product and the second product is determined as the allocation parameter of the transport capacity.

9. A server, characterized in that, The server includes a processor and a memory, the memory storing at least one line of program code, which is loaded and executed by the processor to implement the method for determining target capacity as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one piece of program code, which is loaded and executed by a processor to implement the method for determining target capacity as described in any one of claims 1-8.

Images