WO2021017609A1 - Détermination d'une heure d'arrivée estimée - Google Patents

Détermination d'une heure d'arrivée estimée Download PDF

Info

Publication number
WO2021017609A1
WO2021017609A1 PCT/CN2020/093301 CN2020093301W WO2021017609A1 WO 2021017609 A1 WO2021017609 A1 WO 2021017609A1 CN 2020093301 W CN2020093301 W CN 2020093301W WO 2021017609 A1 WO2021017609 A1 WO 2021017609A1
Authority
WO
WIPO (PCT)
Prior art keywords
waybill
delivery
time
delivery time
capacity
Prior art date
Application number
PCT/CN2020/093301
Other languages
English (en)
Chinese (zh)
Inventor
潘基泽
茹强
周越
闫聪
李梦瑶
Original Assignee
北京三快在线科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 北京三快在线科技有限公司 filed Critical 北京三快在线科技有限公司
Publication of WO2021017609A1 publication Critical patent/WO2021017609A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G06Q10/047—Optimisation of routes or paths, e.g. travelling salesman problem
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/08—Logistics, e.g. warehousing, loading or distribution; Inventory or stock management
    • G06Q10/083—Shipping
    • G06Q10/0835—Relationships between shipper or supplier and carriers
    • G06Q10/08355—Routing methods

Definitions

  • This application relates to the field of computer technology, especially to determining the estimated delivery time.
  • ETA Estimatimated Time of Arrival
  • shippers such as takeaway riders
  • delivery objects such as ordering users
  • the reference time improves the user experience to a certain extent.
  • the ETA estimation method in the prior art usually adopts a combination of logistic regression model and fusion model. For example, input the sample data to be estimated into the logistic regression model to obtain output A, and at the same time according to the prediction method of the sample data Set the attribute value, input the estimated sample data into the corresponding attribute correlation model to obtain the output B, and then perform the weighted fusion of the results of A and B, and then calculate the fusion time according to the rules to make the final output .
  • the embodiments of the present application provide a method for determining the estimated delivery time, including:
  • the pre-trained delivery time estimation model is used to estimate the punctual delivery of the current waybill by each of the shippers at each preset delivery time Probability and capacity capacity coefficient; wherein the capacity capacity coefficient represents the ability index of the dispatcher to carry at least one other waybill when delivering the current waybill;
  • an embodiment of the present application provides a device for determining an estimated delivery time, including:
  • the characteristic acquisition module is used to acquire the characteristics of the current waybill, the characteristics of the dispatcher of each dispatcher, and the characteristics of the carrying waybill of the dispatcher;
  • the multi-point punctual delivery probability estimation module is used to estimate the delivery time estimation model for each of the distributors based on the characteristics of the waybill, the characteristics of the dispatcher, and the characteristics of the baggage waybill.
  • the estimated delivery time selection output module is used to determine the estimated delivery time of the current waybill based on the acquired capacity and demand data, the on-time delivery probability of each preset delivery time, and the capacity coefficient.
  • an embodiment of the present application also discloses an electronic device, including a memory, a processor, and a computer program stored on the memory and capable of running on the processor, and the processor executes the computer program when the computer program is executed.
  • the method for determining the estimated delivery time described in the embodiment of the application is also disclosed.
  • an embodiment of the present application provides a computer-readable storage medium on which a computer program is stored.
  • the program is executed by a processor, the steps of the method for determining the estimated delivery time disclosed in the embodiment of the present application are disclosed.
  • the method for determining the estimated delivery time disclosed in the embodiments of the present application is obtained by obtaining the characteristics of the current waybill, the characteristics of the dispatcher of each dispatcher, and the characteristics of the dispatcher’s carrying waybill; then, based on the characteristics of the waybill and the dispatcher And the characteristics of the baggage waybill, through the pre-trained delivery time estimation model, predict the on-time delivery probability and the capacity coefficient of each of the distributors for the current waybill at each preset delivery time; where The capacity coefficient indicates the ability of the shipper to carry at least one other waybill when delivering the current waybill; finally, according to the obtained capacity and demand data, the on-time delivery probability of each preset delivery time and the The capacity coefficient is used to determine the estimated delivery time of the current waybill. Since the estimated time is dynamically adjusted in combination with specific business goals, the output estimated time can balance user experience and delivery efficiency.
  • Figure 1 is a flow chart of the method for determining the estimated delivery time in the first embodiment of the present application
  • FIG. 2 is a schematic diagram of a network model structure adopted in Embodiment 1 of the present application.
  • FIG. 3 is a schematic diagram of the corresponding relationship between the preset delivery time and the on-time delivery probability of the model output in Embodiment 1 of the present application;
  • FIG. 4 is one of the schematic diagrams of the apparatus for determining the estimated delivery time according to the second embodiment of the present application.
  • FIG. 5 is the second structural diagram of the device for determining the estimated delivery time in the second embodiment of the present application.
  • Fig. 6 shows a block diagram of an electronic device for executing the method according to the present application.
  • Fig. 7 shows a storage unit for holding or carrying program codes for implementing the method according to the present application.
  • An embodiment of the present application discloses a method for determining an estimated delivery time. As shown in FIG. 1, the method includes: step 110 to step 130.
  • Step 110 Obtain the characteristics of the current waybill, the characteristics of the dispatcher of each dispatcher, and the characteristics of the carrying waybill of the dispatcher.
  • the method for determining the estimated delivery time disclosed in the embodiments of the present application is applicable to multiple application fields in the distribution industry such as takeaway and logistics, so as to solve the problem of estimating the delivery time of the waybill.
  • the specific technical solution of the method for determining the estimated delivery time disclosed in this application will be illustrated in combination with application scenarios in the food delivery field.
  • the delivery person described in the embodiment of the application may be a takeaway delivery person, that is, a rider, or a delivery robot.
  • the delivery person is a rider as an example to explain the specific technical solution;
  • the waybill described in the application embodiment corresponds to the delivery task of one waybill.
  • the current waybill when the system generates a new waybill to be delivered, can be generated according to the delivery task of the waybill, and the information of the current waybill includes at least geographic location information, such as the delivery start position and the delivery end position; It can also include information such as the price of the waybill.
  • the time required for the current waybill to be delivered by different riders is determined according to the information of the current waybill and the relevant information of each rider in the system. Finally, the current waybill is further determined according to the target of the delivery task to which or which the current waybill is suitable for Rider delivery, and determine the estimated delivery time when delivered by the corresponding rider.
  • the current waybill characteristics are acquired based on the current waybill information, and the waybill characteristics include any one or more of the waybill price, delivery distance, and geographic location.
  • the waybill information has a fixed format, and by analyzing the waybill information of the current waybill according to the fixed format, the characteristics of the current waybill can be obtained.
  • the relevant information of the dispatcher includes at least the information of the dispatcher himself.
  • the information of the dispatcher himself is expressed through the characteristics of the dispatcher.
  • the characteristics of the dispatcher described in the embodiments of the present application include any one or more of dispatcher level, dispatcher's dispatching ability, and dispatcher's dispatch type.
  • the characteristics of the rider include any one or more of the rider level, the rider's delivery ability, and the rider's delivery type.
  • the characteristics of rider level, rider delivery ability, rider delivery type, etc. are determined by the system through analysis and calculation based on rider's historical delivery data, user comments, rider registration information and other data, and can be obtained through the system interface.
  • the technical means for determining the characteristics of the rider level, the rider's delivery ability, the rider's delivery type and the like through analysis and calculation refer to the prior art, and will not be repeated in the embodiments of this application.
  • the type of rider delivery refers to the type of rider who belongs to urban agency, Meituan self-operated, and crowdsourced delivery.
  • the relevant information of the shipper also includes the information of the waybill carried by the shipper.
  • the information of the air waybill carried by the dispatcher is expressed by the characteristic of the air waybill carried by the dispatcher.
  • the characteristics of the air waybill carried by the shipper include: the air waybill characteristics of the air waybill other than the current waybill carried by the shipper; or, the characteristics of the air waybill carried by the shipper include: the The characteristics of all waybills carried by the shipper.
  • the characteristics of the waybill carried by the rider may include the characteristics of part of the waybill or all the waybills other than the current waybill carried by the rider.
  • the air waybill characteristics of each waybill carried by the rider please refer to the aforementioned specific method of obtaining the air waybill characteristics of the current waybill, which will not be repeated here.
  • a takeaway waybill delivered from location A to location B its characteristics can be expressed as F 1 ⁇ (lng1,lat1),(lng2,lat2),value,Distance ⁇ , where (lng1,lat1) means take Meal location coordinates, (lng2, lat2) represents the delivery location coordinates, value represents the price of the waybill, and Distance represents the delivery distance.
  • the rider characteristics of each rider can be obtained separately, then the rider characteristics of rider n can be expressed as F 2 ⁇ Level n ,ability n ,type n ⁇ , where Level n represents the level of rider n ,Ability n represents the delivery capability of rider n, type n represents the delivery type of rider n, N and n are natural numbers greater than 1, and n ⁇ N. Furthermore, if a rider n is carrying two waybills, they are represented as order1 and order2 respectively. Among them, the waybill feature of the waybill order1 is represented as: order1_F1
  • order2_F1 The waybill feature of order2 is expressed as: order2_F1
  • a set of feature vectors made up of the current waybill characteristics, the rider characteristics of the rider, and the rider's carrying waybill characteristics can be obtained for each rider.
  • the spliced characteristics of rider n can be expressed as F n ⁇ F n1 ,F n2 ,F n3 ⁇ .
  • Step 120 based on the characteristics of the waybill, the characteristics of the shipper, and the characteristics of the baggage waybill, through a pre-trained delivery time estimation model, it is estimated that each of the distributors will respond to the current waybill at each preset delivery time. Probability of on-time delivery and capacity factor.
  • the capacity coefficient represents the ability index of the shipper to be able to carry at least one other waybill when delivering the current waybill.
  • the pre-trained delivery time estimation model is used to estimate the on-time delivery probability and the capacity coefficient corresponding to each delivery time when each distributor delivers the current waybill.
  • the on-time delivery probability is used to indicate the probability of on-time delivery within a specified series of delivery times.
  • the specified series of delivery times can be determined according to specific business requirements, or can be determined according to the delivery distance and delivery time data in the historical waybill data. For example, according to historical waybill data, the longest delivery time is 120 minutes, and the shortest delivery time is 0 minutes, then the specified series of delivery times can be 0 minutes, 1 minute, 2 minutes, ...
  • the specified series of delivery time may be a series of delivery time points such as 0 minutes, 5 minutes, 10 minutes, ... 120 minutes.
  • the delivery time mentioned in the embodiments of the present application refers to the time required by the rider from the start position to the end position of the waybill, for example, the time required from the takeaway rider picking up the meal to the delivery.
  • the pre-trained delivery time estimation model is used to estimate that each of the dispatchers will have the current waybill in each forecast.
  • it also includes: determining the characteristics of each waybill corresponding to each waybill according to the historical waybill data, the characteristics of the dispatcher of the dispatcher who delivered the waybill, and the dispatcher
  • the piggyback air waybill feature is used as the sample data of the training sample corresponding to the corresponding air waybill; and, the sample label of the training sample corresponding to each waybill is determined according to the pickup time of each waybill in the historical air waybill data, the sample label It includes the on-time delivery probability label and the capacity coefficient label corresponding to the preset delivery time; training a delivery time estimation model based on the training sample, wherein the delivery time estimation model is a multi-class model.
  • training samples are constructed based on historical air waybill data to train the delivery time estimation model. For example, for each historical takeaway delivery record, determine the pickup location, delivery location, and waybill price information of the waybill, and encode the waybill feature vector corresponding to the waybill record; at the same time, determine the record of the takeaway delivery record The delivery rider is further obtained, and the rider grade, rider delivery ability, and rider type information of the delivery rider are obtained and encoded to obtain the rider feature vector corresponding to the waybill record.
  • the waybill information of other waybills carried by the rider who delivered the waybill at the same time based on the takeaway delivery record, that is, the waybills of other takeaway waybills that were delivered at the same time when the rider delivered the waybill. information.
  • a certain rider will carry one or more other waybills at the same time, and the waybills carried by the rider at the same time often overlap partly on the delivery route and can share a certain part of the delivery distance.
  • the waybill information of each waybill carried by the rider includes: pick-up location, delivery location, and waybill price information. Then, the above-mentioned characteristic information of each waybill is separately coded and spliced to obtain the characteristic vector of the rider's carrying waybill.
  • the delivery time estimation model trained in the embodiment of the application is a multi-class model, and the task of the multi-class model is to predict the classification results of a piece of test data similar to sample data corresponding to multiple categories.
  • the goal is to input data consisting of the air waybill characteristics, rider characteristics (ie, shipper characteristics), and carrying air waybill characteristics, and the delivery time estimation model will output multiple corresponding classification results , That is, output the on-time delivery probability of each preset delivery time and the capacity coefficient corresponding to the on-time delivery probability.
  • the sample label consists of two parts, which are respectively recorded as the on-time delivery probability label and the capacity coefficient label.
  • the on-time delivery probability label is used to indicate the on-time delivery probability of the training sample at each preset delivery time
  • the capacity coefficient label is used to indicate that the rider corresponding to the training sample is delivering the waybill corresponding to the training sample The ability to carry other air waybills.
  • the step of determining the sample label of the training sample corresponding to each waybill according to the delivery time of each waybill in the historical waybill data includes: for each waybill, according to the The delivery time of the waybill determines the value of the on-time delivery probability label of the training sample corresponding to the waybill, wherein the on-time delivery probability label is used to indicate the on-time delivery of the waybill at each preset delivery time Probability.
  • the on-time delivery probability label is used to indicate the classification results corresponding to each preset delivery time point.
  • the delivery time Take the delivery time from 1 minute to 120 minutes as an example. Assuming that every minute the time changes, there is a delivery time, that is, the delivery time includes 120 delivery times from 1 minute to 120 minutes, and each of them is delivered
  • the sample label is the classification result of on-time delivery probability corresponding to 120 delivery times.
  • the delivery status of a certain waybill at each preset delivery time can be used as the on-time delivery probability label of the training sample corresponding to the waybill. That is, if at a certain point in time, the waybill has been delivered, the classification result of the corresponding category at that time point is 1; and the classification result corresponding to other delivery times is 0.
  • the delivery time of a historical waybill is 45 minutes, that is, the historical waybill has been delivered within 45 minutes of delivery time, that is, the probability of on-time delivery is 100%
  • the training sample corresponding to the historical waybill will be delivered on time
  • the classification result of the probability label corresponding to the 45th category can be set to 1, and the other classification results can be set to 0. That is, for this waybill, the on-time delivery probability tag corresponding to the preset delivery time of 45 minutes can be set to 1, and the on-time delivery probability tag corresponding to the preset delivery time of non-45 minutes can be set to 0 .
  • the step of determining the sample label of the training sample corresponding to each waybill according to the delivery time of each waybill in the historical waybill data further includes: according to the delivery time of each waybill The delivery distance of all the waybills carried by the shipper of the waybill mentioned in the route, the delivery distance of the said waybill is divided, and each segment of the delivery distance corresponds to a segment time; according to the corresponding description of each delivery distance
  • the segment time and the number of waybills sharing the corresponding segment's delivery route determine the delivery time of a single waybill sharing the segment's delivery route; for each waybill sharing this segment's delivery route, according to the delivery route of the waybill
  • the sum of the delivery time of the single air waybill corresponding to each segment of the delivery route and the sum of the segment time lengths of each segment of the delivery route of the air waybill are used to determine the capacity coefficient corresponding to the air waybill;
  • the capacity coefficient corresponding to the waybill is used as the
  • each delivery time that is, the delivery time includes 120 time points from 1 minute to 120 minutes
  • each delivery time Corresponding to a capacity coefficient
  • each waybill data will correspond to 120 capacity capacity coefficients
  • the capacity capacity coefficient corresponding to each delivery time constitutes the capacity capacity coefficient label of a training sample.
  • the following combines a piece of specific historical waybill data of rider A to illustrate the technical solution of determining the capacity coefficient label based on historical waybill data.
  • a specific historical air waybill data of rider A is as follows: take waybill 1 at 9:19:00; take waybill 2 after 380 seconds (ie 9:25:20); after another 900 seconds (ie 9:40:20) Send waybill 1; after another 250 seconds (ie 9:44:30), take waybill 3; after another 130 seconds (ie, 9:46:40), send waybill 2, and take waybill 4 at the same time; after another 770 seconds (Ie 9:59:30) send waybill 3; after 890 seconds (ie 10:14:20) send waybill 4. From this historical waybill data, it can be concluded that rider A is carrying 4 waybills at the same time.
  • the delivery route of this historical waybill data can be divided into multiple delivery routes according to the time of each collection and delivery of the waybill. After the aforementioned historical air waybill data is divided, there will be 6 delivery routes, and the segment durations corresponding to each delivery route are: 380 seconds, 900 seconds, 250 seconds, 130 seconds, 770 seconds and 890 seconds.
  • the delivery route of each waybill may include one segment or multiple consecutive segments.
  • the delivery distance of the waybill 1 includes the first segment (380 seconds) and the second segment (900 seconds) of the delivery distance.
  • the delivery time of a single air waybill sharing the delivery distance is determined according to the segment time corresponding to each segment of the delivery distance included in the delivery distance of each waybill and the number of airway bills sharing the delivery distance.
  • the quotient of the segment duration corresponding to each segment of the delivery route and the number of waybills sharing the corresponding segment's delivery route is determined as the delivery duration of a single airway bill sharing the segment's delivery route. Take the delivery route of waybill 1 as an example.
  • this method it is possible to determine the delivery time of a single air waybill corresponding to each section of the delivery distance included in the delivery distance of each waybill.
  • the sum of the delivery time of a single waybill corresponding to each segment of the delivery route of the waybill can be shared with the segment time corresponding to each segment of the corresponding segment of the delivery route
  • the ratio of the sum is used as the capacity coefficient corresponding to the waybill.
  • the delivery time of a single waybill corresponding to the first segment of the delivery route included in the waybill 1 is 380 seconds and the delivery time of a single waybill corresponding to the second segment of the delivery route included in the waybill 1 is 450 seconds
  • the time obtained by the summation is 830 seconds
  • the capacity coefficient corresponding to the waybill is used as the capacity capacity coefficient label of the training sample corresponding to the waybill.
  • the capacity coefficient of the waybill 1 is both 0.64843750.
  • the historical air waybill data is the data that has actually occurred, so the capacity coefficient of each delivery time of a certain waybill corresponding to the delivery route can be accurately obtained.
  • the capacity coefficient corresponding to the final delivery time of the historical waybill is calculated, and the capacity coefficients of other preset delivery time points are determined by generalization of the model.
  • the delivery time estimation model is further trained based on the training samples.
  • the delivery time estimation model described in the embodiment of the application adopts a multi-target network structure, for example, a classification neural network with two targets as shown in FIG. 2 is adopted, and one of the targets is to output the classification result of the preset delivery time , Another goal is to output the capacity coefficient corresponding to each preset delivery time.
  • the loss functions of the two network branches, the on-time delivery probability loss function 210 and the capacity coefficient loss function 220 are different.
  • the delivery time estimation model further includes: a feature fusion layer 230, which is used to combine the first feature vector output by the waybill feature vectorization module 240 and the shipper feature vectorization module
  • the second feature vector output by 250 is fused to obtain a fusion feature vector of the first preset length, and then the fusion feature vector is input to the loss functions 210 and 220, respectively, to obtain different classification results.
  • the air waybill feature vectorization module 240 is further configured to integrate any one or more of the air waybill price, delivery distance, and geographic location input into the delivery time estimation model into a second forecast. Let the length of the first feature vector.
  • the shipper feature vectorization module 250 is used to fuse the shipper feature of the shipper who delivers the current waybill with the shipper's piggy-backed waybill feature to obtain a second feature vector with a third preset length.
  • the shipper feature vectorization module 250 further includes: a shipper feature fusion sub-module 2501, a shipper feature vectorization sub-module 2502, and a preset number of piggy-back waybill feature vectorization sub-modules 2503.
  • the dispatcher feature vectorization sub-module 2502 is used to fuse any one or more of the characteristic vectors of the dispatcher level, the dispatcher's dispatching ability, and the dispatcher's dispatch type into a dispatcher characteristic of the fourth preset length Vector; each of the piggybacked waybill feature vectorization sub-module 2503 is used to fuse the vector of the waybill feature of a certain waybill carried by the shipper into a first feature vector of a second preset length; the shipper feature fusion
  • the sub-module 2501 is used for fusing the fourth preset length of the shipper feature vector and a preset number of the first feature vector output by the baggage waybill feature vectorization sub-module 2503 into a third preset length of the second feature vector.
  • the loss function and the parameters of each feature fusion module and feature fusion sub-module are continuously adjusted through methods such as back propagation to obtain the optimal model parameters and complete the model training.
  • the specific training process of the delivery time estimation model refers to the prior art, and the specific training process of the model is not limited in the embodiment of this application.
  • the structure of the delivery time estimation model can also be in other forms, for example, the delivery feature fusion submodule 2501 is deleted, and the parameters of the delivery feature vectorization module 250 are adjusted by The shipper feature vectorization module 250 merges the shipper feature vector of the fourth preset length and the preset number of the first feature vector output by the baggage waybill feature vectorization submodule 2503 into a third preset length The second feature vector.
  • a specific network structure can also be used to integrate the features of the waybill, the shipper, and the shipper’s baggage waybill, which can also solve the technical problem to be solved by this application and achieve the same
  • the technical effects are not listed one by one in the embodiments of this application.
  • the delivery time estimation model when estimating the on-time delivery probability of the current waybill, the characteristics of the current waybill, the characteristics of rider A, and the burden of rider A
  • the air waybill features are input to the delivery time estimation model, and the delivery time estimation model will output the on-time delivery probability of rider A at each preset delivery time when the current waybill is delivered by rider A, and rider A's delivery office State the capacity coefficient of the current air waybill at each preset delivery time.
  • each forecast is obtained.
  • the on-time delivery probability and capacity coefficient of the current waybills within different delivery times have been determined when different riders deliver the current waybill.
  • the corresponding table of the estimated rider, the preset delivery time and the on-time delivery probability is shown in Figure 3.
  • the first column on the left represents each preset delivery time; starting from the second column on the left, the data in each column identifies the punctual delivery within different delivery time periods when the corresponding rider delivers the current waybill Probability; starting from the second row from top to bottom, the data in each row represents the on-time delivery probability of different riders in the delivery time period on the left.
  • the value 99% in the bottom left corner of Figure 3 means that within 90 minutes, the probability that rider A will deliver the current waybill on time is 99%; the probability that rider A will deliver the current waybill on time in each time period is the leftmost column .
  • Step 130 Determine the expected delivery time of the current waybill according to the acquired capacity and demand data, the on-time delivery probability of each of the preset delivery times, and the capacity coefficient.
  • the determination of the estimated delivery time of the current waybill based on the acquired capacity and demand data, the on-time delivery probability of each of the preset delivery times, and the capacity capacity coefficient includes : Determine the ratio of demand to capacity; if the ratio meets the preset ratio condition, determine the estimated delivery time of the current waybill according to the on-time delivery probability; if the ratio does not meet the preset ratio condition, then The capacity coefficient determines the estimated delivery time of the current waybill.
  • the role of the estimated delivery time is to adjust the balance between user experience and delivery efficiency. With sufficient capacity, reducing the expected delivery time can improve user experience. When the capacity is tight, extending the estimated delivery time can increase the capacity of the entire system.
  • the demand and capacity are defined to obtain the relationship data between demand and capacity, and then the system is guided to select the appropriate estimated delivery time for output, which effectively balances the delivery efficiency of the user experience and makes the output expected delivery Explainable time is stronger.
  • the capacity and demand data of the distribution system can be obtained, and then the estimated time adjustment is made based on the obtained capacity and demand data.
  • the system defines a requirement as: the sum of the number of waybills within a specified geographic area, and defines a capacity as: the number of deliverers (such as the number of riders) that meet the conditions within the specified geographic area. Since demand and transportation capacity are different in size and magnitude, it is necessary to normalize the demand and transportation capacity into dimensionless numbers.
  • a normalization method in the prior art such as a maximum-minimum method
  • an appropriate estimated delivery time can be selected according to the ratio of demand and capacity.
  • determining the estimated delivery time of the current waybill according to the on-time delivery probability further includes: if the ratio satisfies the preset ratio Condition, the shortest delivery time corresponding to the on-time delivery probability meeting the preset probability condition is determined as the estimated delivery time of the current waybill. If the ratio does not meet the preset ratio condition, determining the estimated delivery time of the current waybill according to the capacity coefficient, further comprising: if the ratio does not meet the preset ratio condition, determining the closest to all The delivery time corresponding to the capacity coefficient of the ratio is used as the estimated delivery time of the current waybill.
  • the ratio of demand to capacity is less than or equal to the preset coefficient threshold, and the preset probability condition is that the on-time delivery probability is greater than 95%.
  • the preset coefficient threshold can be determined according to offline operation experience, for example, set to 80%.
  • the preset ratio condition such as less than 85%
  • the shortest delivery time required when the probability of on-time delivery is greater than 95% will be indicated.
  • the rider information corresponding to the shortest delivery time is also output.
  • the delivery time corresponding to the nearest rider is selected as the estimated delivery time of the current waybill in the rider corresponding to the on-time delivery probability meeting the preset probability condition. At the same time, the rider information of the rider with the closest distance is output.
  • the delivery time and the allocation of riders can be dynamically adjusted, which can further improve the delivery efficiency of the system.
  • the calculated demand and capacity ratio does not meet the preset ratio condition (such as greater than 85%), it indicates that the current capacity is insufficient. For example, if all the riders are in full-load delivery, then the capacity coefficient corresponding to the closest ratio is selected. The delivery time is used as the estimated delivery time of the current waybill, and the rider information corresponding to the capacity coefficient closest to the above ratio is output.
  • the delivery time corresponding to the minimum value of the capacity coefficient greater than the above ratio can also be selected as the The estimated delivery time of the current waybill, and output the rider information corresponding to the minimum capacity coefficient greater than the above ratio.
  • the capacity coefficient reflects the rider’s ability to carry the air waybill at a certain time, and the ratio of capacity to demand reflects the distribution demand of the distribution system.
  • capacity is insufficient, the rider whose capacity coefficient matches the ratio of capacity and demand is selected for order distribution ,
  • the distribution system can operate stably, and at the same time, maximize the user experience.
  • the method for determining the estimated delivery time disclosed in the embodiments of the present application is to obtain the current waybill characteristics, the rider characteristics of each rider, and the rider's carrying waybill characteristics; then, based on the waybill characteristics, the rider characteristics, and the carrying The characteristics of the air waybill, through the pre-trained delivery time estimation model, estimate the on-time delivery probability and capacity coefficient of each rider for the current air waybill at each preset delivery time; wherein, the capacity capacity coefficient represents The rider is able to carry at least one capability index of other waybills when delivering the current waybill; finally, it is determined according to the obtained capacity and demand data, the on-time delivery probability of each preset delivery time, and the capacity coefficient Since the estimated delivery time of the current waybill is dynamically adjusted in combination with specific business goals, the output estimated time can balance user experience and delivery efficiency.
  • the method for determining the estimated delivery time disclosed in the embodiments of the present application is based on a universal delivery time estimation model suitable for multiple scenarios, and based on demand and capacity in real time, it is determined whether to take care of user experience or prefer delivery efficiency, and then , To further select the estimated delivery time of the model and the corresponding rider, so that the output estimated delivery time has a certain ability to explain the business.
  • the model adopts abstract waybill characteristics, rider characteristics, and rider’s carrying waybill characteristics to deliver on time for different riders at each preset delivery time.
  • the probability of arrival is not limited to the specific scenarios of the waybill, and there is no need to retrain the delivery time estimation model for different scenarios, which improves the efficiency of model training and reduces development costs.
  • this application adopts a multi-classification model, and the delivery time estimation model outputs on-time delivery probabilities of multiple preset delivery times, providing a more fine-grained selection range for subsequent adjustment strategies, which can improve the expected delivery of output The accuracy of time.
  • This embodiment discloses a device for determining the estimated delivery time. As shown in FIG. 4, the device includes:
  • the characteristic acquisition module 410 is used to acquire the characteristics of the current waybill, the characteristics of the dispatcher of each dispatcher, and the characteristics of the carrying waybill of the dispatcher;
  • the multi-point punctual delivery probability estimation module 420 is used to estimate the delivery time estimation model for each of the distributors based on the characteristics of the waybill, the characteristics of the shipper, and the characteristics of the baggage waybill.
  • the estimated delivery time selection output module 430 is configured to determine the estimated delivery time of the current waybill based on the acquired capacity and demand data, the on-time delivery probability of each preset delivery time, and the capacity coefficient.
  • the estimated delivery time selection output module 430 is further configured to:
  • the estimated delivery time of the current waybill is determined according to the on-time delivery probability
  • the estimated delivery time of the current waybill is determined according to the capacity coefficient.
  • the step of determining the expected delivery time of the current waybill according to the on-time delivery probability includes:
  • the shortest delivery time corresponding to the on-time delivery probability that satisfies the preset probability condition is determined as the estimated delivery time of the current waybill.
  • the step of determining the estimated delivery time of the current waybill according to the capacity coefficient includes:
  • the delivery time corresponding to the capacity coefficient closest to the ratio is determined as the estimated delivery time of the current waybill.
  • the device further includes:
  • the training sample generating module 440 is used to determine the characteristics of the waybill corresponding to each waybill, the characteristics of the shipper who delivered the waybill, and the characteristics of the baggage waybill of the distributor according to the historical waybill data, as the training sample corresponding to the corresponding waybill And, according to the delivery time of each waybill in the historical waybill data, determine the sample label of each training sample corresponding to the waybill, the sample label including the punctuality corresponding to the preset delivery time Delivery probability label and capacity coefficient label;
  • the estimation model training module 450 is configured to train a delivery time estimation model based on the training samples, wherein the delivery time estimation model is a multi-classification model.
  • the step of determining the sample label of the training sample corresponding to each waybill according to the delivery time of each waybill in the historical waybill data includes:
  • the value of the on-time delivery probability label of the training sample corresponding to the waybill is determined according to the delivery time of the waybill, where the on-time delivery probability label is used to indicate that the waybill is in each pre-order. Set the probability of on-time delivery of the delivery time.
  • the step of determining the sample label of the training sample corresponding to each of the waybills according to the delivery time of each waybill in the historical waybill data further includes:
  • segment the delivery distance of the waybill, and each delivery distance corresponds to a segment time
  • the sum of the delivery time of the individual waybill corresponding to each section of the delivery route of the waybill and each section of the delivery route of the waybill The sum of the segment durations of the journey to determine the capacity coefficient corresponding to the waybill;
  • the capacity coefficient corresponding to the waybill is used as the capacity capacity coefficient label of the training sample corresponding to the waybill.
  • the characteristics of the waybill include any one or more of the price of the waybill, the delivery distance, and the geographic location; the characteristics of the dispatcher include: the rank of the dispatcher, the dispatching ability of the dispatcher, and the dispatcher Any one or more of the delivery types; the characteristics of the waybill carried by the dispatcher include: the characteristics of the waybill other than the current waybill carried by the dispatcher.
  • the device for determining the expected delivery time disclosed in the embodiments of the present application is used to implement the steps of the method for determining the expected delivery time described in the first embodiment of the present application.
  • each module of the device refer to the corresponding steps. I won't repeat it here.
  • the device for determining the estimated delivery time disclosed in the embodiments of the present application obtains the characteristics of the current waybill, the characteristics of the dispatcher of each dispatcher, and the characteristics of the load-bearing waybill of the dispatcher; then, based on the characteristics of the waybill and the dispatcher And the characteristics of the baggage waybill, through the pre-trained delivery time estimation model, predict the on-time delivery probability and the capacity coefficient of each of the distributors for the current waybill at each preset delivery time; where The capacity coefficient indicates the ability of the shipper to carry at least one other waybill when delivering the current waybill; finally, according to the obtained capacity and demand data, the on-time delivery probability of each preset delivery time and the The capacity coefficient is used to determine the estimated delivery time of the current waybill.
  • the output estimated time can balance user experience and delivery efficiency. Based on a universal delivery time estimation model suitable for multiple scenarios, it is determined in real time whether to take care of user experience or prefer delivery efficiency based on demand and capacity, and then further select the estimated delivery time of the model and the corresponding dispatcher , So that the estimated delivery time output has a certain ability to explain the business.
  • the model adopts the abstract waybill feature, the shipper's feature, and the shipper’s baggage waybill feature for different shippers to preset delivery times at each
  • the on-time delivery probability is not limited to the specific scenarios of the waybill, and there is no need to retrain the delivery time estimation model for different scenarios, which improves the efficiency of model training and reduces development costs.
  • this application adopts a multi-classification model, and the delivery time estimation model outputs on-time delivery probabilities of multiple preset delivery times, providing a more fine-grained selection range for subsequent adjustment strategies, which can improve the expected delivery of output The accuracy of time.
  • this application also discloses an electronic device, including a memory, a processor, and a computer program stored on the memory and capable of running on the processor.
  • the processor executes the computer program, the implementation is as in this application.
  • the electronic device may be a PC, a mobile terminal, a personal digital assistant, a tablet computer, etc.
  • the application also discloses a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of the method for determining the estimated delivery time as described in the first embodiment of the application are realized.
  • the device embodiments described above are merely illustrative.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in One place, or it can be distributed to multiple network units. Some or all of the modules may be selected according to actual needs to achieve the objectives of the solutions of the embodiments. Those of ordinary skill in the art can understand and implement it without creative work.
  • each implementation manner can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware.
  • the above technical solutions can be embodied in the form of software products, which can be stored in computer-readable storage media, such as ROM/RAM, magnetic A disc, an optical disc, etc., include a number of instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiment.
  • a microprocessor or a digital signal processor may be used in practice to implement some or all of the functions of some or all of the components in the electronic device according to the embodiments of the present application.
  • DSP digital signal processor
  • it may be implemented as a device or device program (for example, a computer program and a computer program product) for executing part or all of the methods described herein.
  • Such a program for realizing the present application may be stored on a computer-readable medium, or may have the form of one or more signals.
  • signals can be downloaded from Internet websites, or provided on carrier signals, or provided in any other form.
  • FIG. 6 shows an electronic device that can implement the method according to the present application.
  • the electronic device traditionally includes a processor 620 and a computer program product in the form of a memory 610 or a computer-readable medium.
  • the memory 610 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read Only Memory), EPROM, hard disk, or ROM.
  • the memory 610 has a storage space 6101 for executing the program code 6102 of any method step in the above method.
  • the storage space 6101 for program codes may include various program codes 6102 for implementing various steps in the above method. These program codes can be read out from or written into one or more computer program products.
  • These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards or floppy disks.
  • Such a computer program product is usually a portable or fixed storage unit as described with reference to FIG. 7.
  • the storage unit may have storage segments, storage spaces, etc. arranged similarly to the storage 620 in the electronic device of FIG. 6.
  • the program code can be compressed in an appropriate form, for example.
  • the storage unit includes computer-readable code 6102', that is, code that can be read by a processor such as 610.
  • the computer-readable code runs on an electronic device, it causes the electronic device to perform the determination described above. The steps in the method of estimated delivery time.
  • any reference signs placed between parentheses should not be constructed as a limitation to the claims.
  • the word “comprising” does not exclude the presence of elements or steps not listed in the claims.
  • the word “a” or “an” preceding an element does not exclude the presence of multiple such elements.
  • the application can be implemented by means of hardware including several different elements and by means of a suitably programmed computer. In the unit claims enumerating several devices, several of these devices may be embodied by the same hardware item.
  • the use of the words first, second, and third, etc. do not indicate any order. These words can be interpreted as names.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Strategic Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Molecular Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Tourism & Hospitality (AREA)
  • Operations Research (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Health & Medical Sciences (AREA)
  • General Business, Economics & Management (AREA)
  • Computing Systems (AREA)
  • Development Economics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Game Theory and Decision Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

L'invention concerne un procédé permettant de déterminer une heure d'arrivée estimée se rapportant au domaine technique des ordinateurs. Le procédé de détermination de l'heure d'arrivée estimée consiste à : obtenir les caractéristiques de commande de la commande actuelle, les caractéristiques de distributeur de chaque distributeur et les caractéristiques de commande acceptées du distributeur (110) ; estimer, au moyen d'un modèle d'estimation d'heure d'arrivée pré-appris, la probabilité d'arrivée à temps et le coefficient de capacité de transport de chaque distributeur pour la commande actuelle à chaque heure d'arrivée prédéfinie d'après les caractéristiques de commande, les caractéristiques de distributeur et les caractéristiques de commande acceptées (120) ; et déterminer l'heure d'arrivée estimée de la commande actuelle en fonction des données de capacité de transport et d'exigences obtenues, de la probabilité d'arrivée à temps à chaque heure d'arrivée prédéfinie et du coefficient de capacité de transport. Comme l'heure estimée est ajustée de manière dynamique en combinaison avec des objectifs de service spécifiques, la génération de l'heure estimée permet d'équilibrer l'expérience de l'utilisateur et l'efficacité de distribution.
PCT/CN2020/093301 2019-07-30 2020-05-29 Détermination d'une heure d'arrivée estimée WO2021017609A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201910696204.6 2019-07-30
CN201910696204.6A CN110543968B (zh) 2019-07-30 2019-07-30 预计送达时间确定方法、装置、电子设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021017609A1 true WO2021017609A1 (fr) 2021-02-04

Family

ID=68709924

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/093301 WO2021017609A1 (fr) 2019-07-30 2020-05-29 Détermination d'une heure d'arrivée estimée

Country Status (2)

Country Link
CN (1) CN110543968B (fr)
WO (1) WO2021017609A1 (fr)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112926922A (zh) * 2021-03-18 2021-06-08 拉扎斯网络科技（上海）有限公司 一种责任判定系统、方法、装置、存储介质以及电子设备
CN113159281A (zh) * 2021-03-25 2021-07-23 拉扎斯网络科技（上海）有限公司 数据处理方法和数据处理装置
CN113159685A (zh) * 2021-04-26 2021-07-23 拉扎斯网络科技（上海）有限公司 配送时长更新方法、装置、电子设备、介质及程序产品
CN113283830A (zh) * 2021-04-29 2021-08-20 北京京东振世信息技术有限公司 运单信息序列生成方法、装置、设备和计算机可读介质
US20220414592A1 (en) * 2021-06-25 2022-12-29 Maplebear Inc.(dba Instacart) Determining estimated delivery time of items obtained from a warehouse for users of an online concierge system to reduce probabilities of delivery after the estimated delivery time

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110543968B (zh) * 2019-07-30 2022-04-12 北京三快在线科技有限公司 预计送达时间确定方法、装置、电子设备及存储介质
CN113222487B (zh) * 2020-01-21 2023-04-18 北京三快在线科技有限公司 调度路径生成方法，装置，存储介质及电子设备
CN113159659A (zh) * 2020-01-22 2021-07-23 北京京东振世信息技术有限公司 更新货单时效的方法、装置、设备和计算机可读介质
CN113313439B (zh) * 2020-02-26 2024-04-05 北京京东振世信息技术有限公司 一种计算妥投时长的方法和装置
CN111523802B (zh) * 2020-04-22 2023-08-08 北京京东振世信息技术有限公司 送达时间应答方法、装置、设备及介质
CN111860906B (zh) * 2020-04-24 2024-04-26 北京嘀嘀无限科技发展有限公司 一种确定预估到达时间的方法和系统
CN111598487B (zh) * 2020-06-22 2024-03-15 拉扎斯网络科技（上海）有限公司 数据处理和模型训练方法、装置、电子设备和存储介质
CN112116151A (zh) * 2020-09-17 2020-12-22 北京嘀嘀无限科技发展有限公司 一种接驾时间预估方法和系统
CN114330797A (zh) * 2020-09-27 2022-04-12 北京三快在线科技有限公司 配送时长预测方法、装置、存储介质和电子设备
CN112668964A (zh) * 2020-12-23 2021-04-16 北京京东振世信息技术有限公司 运输凭证路由方法、装置、设备及计算机可读存储介质
CN113034075A (zh) * 2021-03-29 2021-06-25 上海寻梦信息技术有限公司 物流运单时效推送方法、系统、设备及存储介质
CN113011665A (zh) * 2021-03-29 2021-06-22 上海寻梦信息技术有限公司 物流时效预测方法、装置、设备及存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107766962A (zh) * 2016-08-17 2018-03-06 北京京东尚科信息技术有限公司 一种推荐最优配送时段的方法和装置
CN108122042A (zh) * 2016-11-28 2018-06-05 北京小度信息科技有限公司 配送时间预估方法及装置
CN108491951A (zh) * 2018-01-25 2018-09-04 北京三快在线科技有限公司 一种外卖配送时间的预测方法、装置及电子设备
US20190130260A1 (en) * 2017-10-30 2019-05-02 DoorDash, Inc. System for dynamic estimated time of arrival predictive updates
CN109726843A (zh) * 2017-10-30 2019-05-07 阿里巴巴集团控股有限公司 配送数据预测的方法、装置和终端
CN110543968A (zh) * 2019-07-30 2019-12-06 北京三快在线科技有限公司 预计送达时间确定方法、装置、电子设备及存储介质

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20190130320A1 (en) * 2017-11-02 2019-05-02 Uber Technologies, Inc. Network computer system to implement dynamic provisioning for fulfilling delivery orders
CN109284956B (zh) * 2018-08-10 2020-10-02 北京三快在线科技有限公司 一种任务时长确定方法及装置和电子设备
CN109993367A (zh) * 2019-04-04 2019-07-09 拉扎斯网络科技（上海）有限公司 配送时长的估计方法、估计装置、存储介质和电子设备

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107766962A (zh) * 2016-08-17 2018-03-06 北京京东尚科信息技术有限公司 一种推荐最优配送时段的方法和装置
CN108122042A (zh) * 2016-11-28 2018-06-05 北京小度信息科技有限公司 配送时间预估方法及装置
US20190130260A1 (en) * 2017-10-30 2019-05-02 DoorDash, Inc. System for dynamic estimated time of arrival predictive updates
CN109726843A (zh) * 2017-10-30 2019-05-07 阿里巴巴集团控股有限公司 配送数据预测的方法、装置和终端
CN108491951A (zh) * 2018-01-25 2018-09-04 北京三快在线科技有限公司 一种外卖配送时间的预测方法、装置及电子设备
CN110543968A (zh) * 2019-07-30 2019-12-06 北京三快在线科技有限公司 预计送达时间确定方法、装置、电子设备及存储介质

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112926922A (zh) * 2021-03-18 2021-06-08 拉扎斯网络科技（上海）有限公司 一种责任判定系统、方法、装置、存储介质以及电子设备
CN112926922B (zh) * 2021-03-18 2023-07-18 拉扎斯网络科技（上海）有限公司 一种责任判定系统、方法、装置、存储介质以及电子设备
CN113159281A (zh) * 2021-03-25 2021-07-23 拉扎斯网络科技（上海）有限公司 数据处理方法和数据处理装置
CN113159685A (zh) * 2021-04-26 2021-07-23 拉扎斯网络科技（上海）有限公司 配送时长更新方法、装置、电子设备、介质及程序产品
CN113283830A (zh) * 2021-04-29 2021-08-20 北京京东振世信息技术有限公司 运单信息序列生成方法、装置、设备和计算机可读介质
CN113283830B (zh) * 2021-04-29 2024-04-09 北京京东振世信息技术有限公司 运单信息序列生成方法、装置、设备和计算机可读介质
US20220414592A1 (en) * 2021-06-25 2022-12-29 Maplebear Inc.(dba Instacart) Determining estimated delivery time of items obtained from a warehouse for users of an online concierge system to reduce probabilities of delivery after the estimated delivery time
WO2022271287A1 (fr) * 2021-06-25 2022-12-29 Maplebear Inc. (Dba Instacart) Détermination de l'heure de livraison estimée d'articles obtenus à partir d'un entrepôt pour des utilisateurs d'un système de conciergerie en ligne pour réduire les probabilités de livraison après l'heure de livraison estimée
US11755987B2 (en) * 2021-06-25 2023-09-12 Maplebear Inc. Determining estimated delivery time of items obtained from a warehouse for users of an online concierge system to reduce probabilities of delivery after the estimated delivery time

Also Published As

Publication number Publication date
CN110543968B (zh) 2022-04-12
CN110543968A (zh) 2019-12-06

Similar Documents

Publication Publication Date Title
WO2021017609A1 (fr) Détermination d'une heure d'arrivée estimée
US8788439B2 (en) Instance weighted learning machine learning model
Shang et al. Distribution network redesign for marketing competitiveness
US20190251609A1 (en) Commodity demand prediction system, commodity demand prediction method, and commodity demand prediction program
CN109426983A (zh) 闪购活动自动生成方法及装置、存储介质、电子设备
CN110503264A (zh) 一种货源排序方法、装置、设备和存储介质
CN109767333A (zh) 选基方法、装置、电子设备及计算机可读存储介质
Ordoobadi Inclusion of risk in evaluation of advanced technologies
LU503730B1 (en) Sequence recommendation method and system based on the coupling relationship between item attributes and time sequence patterns
Kawakami et al. Seasonal Inventory Management Model for Raw Materials in Steel Industry
CN111353728A (zh) 一种风险分析方法和系统
CN115439180A (zh) 一种目标对象确定方法、装置、电子设备及存储介质
CN113689237A (zh) 待投放媒体资源、媒体资源处理模型的确定方法及装置
JP7012892B1 (ja) 情報処理装置、情報処理方法および情報処理プログラム
Özkan et al. A simulation model for evaluating the cargo transfer alternatives in liquid cargo terminals
Lucic et al. Performance of the autoregressive integrated moving average model with exogenous variables statistical model on the intraday market for the Denmark-West bidding area
CN112396498A (zh) 一种商品推销方法、装置、设备及存储介质
CN112632404A (zh) 基于多粒度自注意力的下一个兴趣点推荐方法
CN110781929A (zh) 信用预测模型的训练方法、预测方法及装置、介质和设备
Viellechner The new era of predictive analytics in container shipping and air cargo
CN107886357A (zh) 基于用户行为数据判定内容价值的方法和系统
US20210256509A1 (en) Planning currency exchange transactions based on international travel budgeting information
CN114970923A (zh) 配送订单合包派发方法、装置、电子设备
JP7090785B1 (ja) 情報処理装置、情報処理方法、及び情報処理プログラム
Kashtanov et al. Using Triple Exponential Smoothing and Autoregressive Models to Mining Equipment Details Sales Forecast

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20847202

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20847202

Country of ref document: EP

Kind code of ref document: A1