Distribution time prediction method and device

Through the PID adjustment method, the initial prediction results are adjusted using the time difference signal in the historical period, which solves the problem of inaccurate delivery time prediction in the logistics industry, and realizes accurate delivery time prediction in local areas, improving the satisfaction of users and riders degree and efficiency.

CN118982115APending Publication Date: 2024-11-19SHANGHAI FENGNIAO JIPEI INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411064155.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-02
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The prior art cannot accurately predict the delivery time of local areas in the logistics industry, resulting in inaccurate prediction of delivery time, affecting user satisfaction and rider efficiency.

Method used

The PID adjustment method is used to obtain the historical predicted time difference and the target predicted time difference within the historical period, generate the target difference signal, and adjust the initial prediction results to determine the accurate delivery time.

Benefits of technology

Real-time, fast and accurate prediction of local area order delivery time is achieved, user experience and rider efficiency are improved, and prediction error fluctuations are reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118982115A_ABST
    Figure CN118982115A_ABST
Patent Text Reader

Abstract

The embodiment of the invention discloses a distribution time prediction method and device. The method comprises the steps of obtaining an initial prediction result of a current order; a first difference value signal, a second difference value signal and a third difference value signal are obtained, a target difference value signal is generated according to the first difference value signal, the second difference value signal and the third difference value signal, and the target difference value signal is a predicted difference value between predicted delivery time and actual delivery time of the current order; and adjusting the initial prediction result of the current order according to the target difference signal, and determining the predicted delivery time of the current order. Therefore, in the embodiment of the invention, the distribution time of the order in the local area can be predicted in real time by adopting the mode, and the distribution time of the order can be rapidly and accurately predicted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a delivery time prediction method and device. Background Art

[0002] As the pace of life continues to accelerate, real-time prediction of delivery time plays a vital role in the logistics industry. Accurate delivery time prediction is not only related to user satisfaction, but also directly affects the delivery efficiency of the rider. In the existing technology, when predicting delivery time, the overall regional prediction method is generally adopted, and the delivery time of the local area cannot be predicted, or the predicted delivery time is too long to achieve real-time prediction. Therefore, how to quickly and accurately predict the delivery time is a problem that needs to be solved urgently. Summary of the invention

[0003] In view of this, an embodiment of the present invention provides a delivery time prediction method, device, electronic device and readable storage medium, which adopts PID adjustment to make real-time prediction of the delivery time of orders in a local area, and can quickly and accurately predict the delivery time.

[0004] In a first aspect, an embodiment of the present invention provides a method for predicting delivery time, the method comprising:

[0005] Obtain an initial prediction result for the current order; obtain a first difference signal, a second difference signal, and a third difference signal, wherein the first difference signal represents a first difference between a historical prediction time difference and a target prediction time difference in a first historical period, the second difference signal represents a sum of at least one second difference between a historical prediction time difference and a target prediction time difference in a second historical period, and the third difference signal represents a difference between a historical prediction time difference in the first historical period and a historical prediction time difference in a third historical period, wherein the first historical period is a unit period before the current moment, the second historical period is a period composed of a plurality of unit periods before the current moment, and the third historical period is a unit period before the first historical period, and the prediction time difference is a difference between a predicted delivery time and an actual delivery time in a corresponding time period; generate a target difference signal according to the first difference signal, the second difference signal, and the third difference signal, wherein the target difference signal is a predicted difference between a predicted delivery time and an actual delivery time of the current order; adjust the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order.

[0006] Optionally, the acquiring the first difference signal, the second difference signal and the third difference signal comprises:

[0007] Acquire a candidate first difference signal, wherein the candidate first difference signal represents a plurality of candidate first differences between historical prediction time differences and target prediction time differences of a plurality of orders in a first historical period;

[0008] Perform outlier filtering on the plurality of candidate first difference values ​​by using a signal filter to determine a plurality of filtered candidate first difference values;

[0009] An average value of the plurality of first difference candidates after filtering is determined as the first difference signal.

[0010] Optionally, the acquiring the first difference signal, the second difference signal and the third difference signal comprises:

[0011] Acquire a candidate second difference signal, wherein the candidate second difference signal represents a plurality of candidate second differences between historical prediction time differences and target prediction time differences of a plurality of orders in a second historical period;

[0012] Perform outlier filtering on the plurality of candidate second difference values ​​by using a signal filter to determine a plurality of filtered candidate second difference values;

[0013] Determine at least one second difference value according to the filtered multiple candidate second difference values, wherein each unit period included in the second historical period corresponds to a second difference value, and the second difference value is an average value of multiple candidate second difference values ​​in each unit period;

[0014] The sum of the at least one second difference value is determined as the second difference signal.

[0015] Optionally, the acquiring the first difference signal, the second difference signal and the third difference signal comprises:

[0016] Obtaining historical prediction time differences of multiple orders in a first historical period and historical prediction time differences of multiple orders in a third historical period;

[0017] Performing outlier filtering on the historical prediction time differences of the plurality of orders in the first historical period and the historical prediction time differences of the plurality of orders in the third historical period through a signal filter, and determining the filtered historical prediction time differences of the plurality of orders in the first historical period and the filtered historical prediction time differences of the plurality of orders in the third historical period;

[0018] Determine a first average value of the historical prediction time differences of the plurality of orders in the first filtered historical period, and a second average value of the historical prediction time differences of the plurality of orders in the third filtered historical period;

[0019] A difference between the first average value and the second average value is determined as the third difference signal.

[0020] Optionally, generating a target difference signal according to the first difference signal, the second difference signal and the third difference signal specifically includes:

[0021] determining a first product of a first weight and the target difference signal, a second product of a second weight and the second difference signal, and a third product of a third weight and the third difference signal;

[0022] A sum of the first product, the second product and the third product is determined as the target difference signal.

[0023] Optionally, the obtaining of the initial prediction result of the current order specifically includes:

[0024] Get the order characteristics of the current order;

[0025] The order features are input into a pre-trained machine learning model to determine an initial prediction result for the current order.

[0026] Optionally, adjusting the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order specifically includes:

[0027] The target difference signal is superimposed on the initial prediction result of the current order to generate a predicted delivery time for the current order.

[0028] Optionally, the method further comprises:

[0029] The predicted delivery time is sent to the user terminal.

[0030] Optionally, the first difference signal represents a first difference between a historical prediction time difference in a set area for a first historical period and a target prediction time difference, the second difference signal represents a sum of at least one second difference between a historical prediction time difference in a set area for a second historical period and a target prediction time difference, and the third difference signal represents a difference between a historical prediction time difference in the first historical period in the set area and a historical prediction time difference in the third historical period in the set area, and the set area is the delivery area to which the current order belongs.

[0031] Optionally, the first weight, the second weight and the third weight are set according to a scenario, wherein the scenario includes a set time period and a set area, wherein the set time period is a unit time period in which the current order is located, and the set area is a delivery area to which the current order belongs.

[0032] Optionally, the method further comprises:

[0033] The first difference signal, the second difference signal and the third difference signal are updated every set period.

[0034] In a second aspect, an embodiment of the present invention provides a delivery time prediction device, the device comprising:

[0035] An acquisition unit, used to acquire the initial prediction result of the current order;

[0036] The acquisition unit is further used to acquire a first difference signal, a second difference signal and a third difference signal, wherein the first difference signal represents a first difference between a historical prediction time difference of a first historical period and a target prediction time difference, the second difference signal represents a second difference between a historical prediction time difference of a second historical period and a target prediction time difference, and the third difference signal represents a difference between a historical prediction time difference of the first historical period and a historical prediction time difference of a third historical period, the first historical period is a period before the current moment, the second historical period is a plurality of periods before the current moment, the third historical period is a period before the first historical period, and the prediction time difference is a difference between a predicted delivery time and an actual delivery time;

[0037] a generating unit, configured to perform PID adjustment on the first difference signal, the second difference signal and the third difference signal to generate a target difference signal, wherein the target difference signal is a predicted difference between a predicted delivery time and an actual delivery time of the current order;

[0038] An adjustment unit is used to adjust the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order.

[0039] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method as described in the first aspect and any one of the first aspects.

[0040] In a fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method as described in the first aspect and any one of the first aspects is implemented.

[0041] The embodiment of the present invention obtains the initial prediction result of the current order; obtains a first difference signal, a second difference signal and a third difference signal, wherein the first difference signal represents the first difference between the historical prediction time difference of the first historical period and the target prediction time difference, the second difference signal represents the sum of at least one second difference between the historical prediction time difference of the second historical period and the target prediction time difference, and the third difference signal represents the difference between the historical prediction time difference of the first historical period and the historical prediction time difference of the third historical period, the first historical period is the unit period before the current moment, the second historical period is the period composed of the first multiple unit periods before the current moment, the third historical period is the unit period before the first historical period, and the predicted time difference is the difference between the predicted delivery time and the actual delivery time in the corresponding time period; generates a target difference signal according to the first difference signal, the second difference signal and the third difference signal, wherein the target difference signal is the predicted difference between the predicted delivery time and the actual delivery time of the current order; adjusts the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order. Therefore, in the embodiment of the present invention, the above method can be used to make real-time predictions on the delivery time of orders in a local area, and the delivery time can be predicted quickly and accurately. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0043] Figure 1 is a schematic diagram of a system architecture of an embodiment of the present invention;

[0044] Figure 2 is a schematic diagram of a regulating module according to an embodiment of the present invention;

[0045] Figure 3 is a flow chart of delivery time prediction in an optional implementation of the first embodiment of the present invention;

[0046] Figure 4 is a flow chart of a method for obtaining a first difference signal according to a first embodiment of the present invention;

[0047] Figure 5 is a flow chart of a method for obtaining a second difference signal according to a first embodiment of the present invention;

[0048] Figure 6 is a flow chart of a method for obtaining a third difference signal according to a first embodiment of the present invention;

[0049] Figure 7 is a flow chart of delivery time prediction in another optional implementation of the first embodiment of the present invention;

[0050] Figure 8 is a schematic diagram of a delivery time prediction device according to a second embodiment of the present invention;

[0051] Fig. 9 is a schematic diagram of an electronic device according to a third embodiment of the present invention. DETAILED DESCRIPTION

[0052] The present application is described below based on embodiments, but the present application is not limited to these embodiments. In the detailed description of the present application below, some specific details are described in detail. It is possible for those skilled in the art to fully understand the present application without the description of these details. In order to avoid confusing the essence of the present application, known methods, processes, flows, components and circuits are not described in detail.

[0053] In addition, persons of ordinary skill in the art will appreciate that the drawings provided herein are for illustration purposes and are not necessarily drawn to scale.

[0054] Unless the context clearly requires otherwise, the words "include", "comprising" and similar words throughout the application should be interpreted as including rather than exclusive or exhaustive; that is, the meaning is "including but not limited to".

[0055] In the description of this application, it should be understood that the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance. In addition, in the description of this application, unless otherwise specified, the meaning of "plurality" is two or more.

[0056] In the embodiments of the present invention, the application scenario of delivery of takeaway goods (including electronic devices, meals, clothing, etc.) is used as an example for description. However, those skilled in the art will readily appreciate that the method of the embodiments of the present invention is also applicable to other application scenarios of transporting goods, such as moving within the same city (or across cities).

[0057] Figure 1 It is a schematic diagram of the system architecture of an embodiment of the present invention. Figure 1 The hardware system architecture shown may include a user terminal and a server, wherein the server includes a prediction module and an adjustment module. Figure 1 A server 11 and a user terminal 12 are taken as an example for description. The server 11 includes a prediction module 1101 and an adjustment module 1102 .

[0058] In this embodiment, the prediction module 1101 in the server 11 is used to generate an initial prediction result for the current order, that is, the predicted delivery time of the current order before the prediction error is calculated. The prediction module can use a machine learning model to predict the initial prediction result; the adjustment module 1102 in the server 11 is used to determine the predicted difference between the predicted delivery time and the actual delivery time of the current order, and then adjust the initial prediction result according to the predicted difference to generate the predicted delivery time of the current order, and send the predicted delivery time to the user terminal 12, wherein the user terminal 12 can be multiple; the user terminal 12 receives the predicted delivery time and displays the predicted delivery time to the user, and the user terminal 12 is also used to send the predicted delivery time to the prediction module 1101 of the server 11 for the delivery time prediction of the next cycle.

[0059] In the process of predicting the delivery time, it is necessary to ensure the stability of the predicted delivery time, that is, the predicted difference between the predicted delivery time and the actual delivery time is stable within a certain range. If the predicted difference is unstable, the following two situations will occur: Situation 1: The predicted delivery time is too long, which will affect the number of orders placed by users. For example, if the predicted delivery time is 1 hour and 30 minutes, users may not place orders due to the long waiting time, which affects the user's satisfaction with the application; Situation 2: If the predicted time is too short, it will cause difficulties for the rider to deliver the order. For example, the predicted delivery time is 20 minutes, but due to the distance, the pressure of delivering the food, etc., the rider cannot deliver the order within 20 minutes, which affects the rider's delivery efficiency and the user's experience. In the prior art, when predicting the delivery time, the overall regional prediction method is generally adopted, and the delivery time of the local area cannot be predicted, or the predicted delivery time is too long and real-time prediction cannot be achieved; therefore, how to quickly and accurately predict the delivery time is a problem that needs to be solved.

[0060] In this embodiment, in order to solve the above problems, a method of adjustment based on dynamic control theory is proposed, that is, the predicted delivery time is determined by using the proportional integration differential (PID) adjustment method. Specifically, the PID controller is divided into proportional P control, integral I control and differential D control. Among them, the P proportional control has a fast response speed and can act on the output quickly. The larger the P, the stronger the proportional effect, the faster the dynamic response, and the stronger the ability to eliminate errors. However, since the actual control system has inertia, a too strong proportional effect will cause system oscillation instability; the I integral control has high accuracy and can eliminate past accumulated errors, but it will affect the stability of the control system; the D differential control has high stability and has a leading control effect; therefore, the PID controller requires P, I, and D to be adjusted together to realize a PID controller with simple algorithm, good robustness, and high reliability. Figure 2 Schematic diagram of an adjustment module according to an embodiment of the present invention. Figure 2 The regulation module shown may include a collector 21 , a signal filter 22 and a PID controller 23 .

[0061] Figure 3 FIG. 1 is a flow chart of a method for predicting delivery time according to a first embodiment of the present invention. Figure 3 As shown, the method of this embodiment includes the following steps:

[0062] Step S301: Obtain the initial prediction result of the current order.

[0063] Specifically, the order features of the current order are obtained, and the order features are input into a pre-trained machine learning model to generate an initial prediction result for the current order; wherein the order features may be merchant features, capacity features, etc., the merchant features may be the number of orders received by the merchant, the meal delivery time of the merchant, etc., and the capacity features may be the number of riders in the area where the order is located, the rider delivery efficiency, etc., which are determined based on actual conditions.

[0064] In this embodiment, the current order is an order within a set time period and a set area. For example, the set time period is the lunch period, and the set area is business district 1; or, the set time period is the dinner period, and the set area is business district 2. This is only an example description and will be determined based on actual conditions.

[0065] Step S302: Acquire a first difference signal, a second difference signal, and a third difference signal.

[0066] Specifically, the first difference signal represents a first difference between a historical predicted time difference in a set area for a first historical period and a target predicted time difference, the second difference signal represents a sum of at least one second difference between a historical predicted time difference in a set area for a second historical period and a target predicted time difference, and the third difference signal represents a difference between a historical predicted time difference in a set area for the first historical period and a historical predicted time difference in the third historical period, the set area being the delivery area to which the current order belongs, the first historical period being a unit period before the current moment, the second historical period being a period composed of multiple unit periods before the current moment, the third historical period being a unit period before the first historical period, and the predicted time difference being the difference between the predicted delivery time and the actual delivery time in the corresponding time period.

[0067] In this embodiment, the unit period is the minimum time unit for obtaining the difference signal, and the first historical period, the second historical period and the third historical period are all composed of the unit period; the first historical period is composed of one unit period, and the first historical period is one unit period before the current moment corresponding to the current order; the second historical period is composed of multiple unit periods, and the second historical period is multiple unit periods before the current moment corresponding to the current order; the third historical period is composed of one unit period, and the third historical period is the second to last unit period before the current moment corresponding to the current order.

[0068] In this embodiment, the specific process of obtaining the first difference signal is as follows: Figure 4 As shown, Figure 4 FIG. 1 is a flow chart of a method for obtaining a first difference signal according to a first embodiment of the present invention. Figure 4 As shown, the method of this embodiment includes the following steps:

[0069] Step S401, obtaining a candidate first difference signal, wherein the candidate first difference signal represents a plurality of candidate first differences between a historical prediction time difference and a target prediction time difference of a plurality of orders in a first historical period.

[0070] For example, the candidate first difference signal is a plurality of candidate first differences between the historical predicted time difference TGAP of multiple orders within N hours before the corresponding moment of the current order and the target time difference, and the predicted time difference is the difference between the predicted delivery time and the actual delivery time in the previous N hours, wherein N is a preset value, which can be an integer or a decimal, and is determined according to the user's needs. Assuming that the user needs to predict the delivery time at a fine granularity, N can be set to 1 hour, that is, with 1 hour as a cycle, obtain multiple candidate first differences between the historical predicted time difference TGAP and the target time difference one hour before the current order occurs; assuming that there are 10 orders, the candidate first differences are 10, namely 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes and 20 minutes.

[0071] Step S402: filtering the plurality of candidate first difference values ​​for outliers using a signal filter to determine a plurality of filtered candidate first difference values.

[0072] Specifically, the outliers in the candidate first difference values ​​are filtered. For example, among the 10 candidate first difference values ​​of 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes and 20 minutes, 20 minutes is an outlier. The 20 minutes are deleted, and the multiple candidate first difference values ​​after filtering are determined, namely 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes and 4 minutes. This is only an exemplary description and will be determined according to actual conditions.

[0073] In this embodiment, the abnormal value may be an abnormal value from a business perspective in addition to the outlier value described above, for example, a candidate first difference value corresponding to an uncompleted order, etc.

[0074] In this example, based on the scale of the order volume in the delivery area, the capacity structure, and the supply and demand differences in different time periods, a funnel model is used to fit a signal filter, wherein the funnel model is a judgment model used to filter abnormal data; it is easy to understand that the signal filter mainly filters the difference signals corresponding to order samples such as those with extremely small orders and extremely difficult delivery (for example, extreme scenarios such as heavy snow and thunderstorms) to ensure stable delivery time control.

[0075] Step S403: Determine the average value of the filtered plurality of candidate first difference values ​​as the first difference signal.

[0076] In this embodiment, assuming that the multiple candidate first difference values ​​after filtering are 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes and 4 minutes, the average value of the multiple candidate first difference values ​​is (5+3+5+4+3+2+6+4+4) minutes / 9=4 minutes, then the first difference signal is 4 minutes. This is only an exemplary explanation and will be determined based on actual conditions.

[0077] In this embodiment, the specific process of obtaining the second difference signal is as follows: Figure 5 As shown, Figure 5 FIG. 1 is a flow chart of a method for obtaining a second difference signal according to a first embodiment of the present invention. Figure 5 As shown, the method of this embodiment includes the following steps:

[0078] Step S501, obtaining a candidate second difference signal, wherein the candidate first difference signal represents a plurality of candidate second differences between a historical prediction time difference and a target prediction time difference of a plurality of orders in a second historical period.

[0079] For example, the candidate second difference signal is a plurality of candidate second differences between the historical predicted time difference TGAP of multiple orders within the M hours before the corresponding moment of the current order and the target time difference, and the predicted time difference is the difference between the predicted delivery time and the actual delivery time in the previous M hours, wherein the value of M is greater than the value of N, and M is a preset value, which can be an integer or a decimal, and is determined specifically according to user needs. Assuming that N can be set to 3 hours, a plurality of candidate second differences between the historical predicted time difference TGAP and the target time difference 3 hours before the current order is sent are obtained; assuming that there are 20 orders, the candidate second differences are 10, namely 5 minutes, 4 minutes, 5 minutes, 5 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes, 18 minutes, 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 3 minutes, 6 minutes, 4 minutes, 4 minutes and 0 minutes.

[0080] Step S502: Filter the plurality of candidate second difference values ​​for outliers using a signal filter to determine a plurality of filtered candidate second difference values.

[0081] Specifically, the outliers in the candidate second difference values ​​are filtered. For example, among the 10 candidate second difference values ​​of 5 minutes, 4 minutes, 5 minutes, 5 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes, 18 minutes, 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 3 minutes, 6 minutes, 4 minutes, 4 minutes and 0 minutes, 10 minutes is an outlier and 0 minutes is a business outlier that does not complete the order normally. 18 minutes and 0 minutes are deleted to determine a plurality of filtered candidate second difference values, namely 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes, 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 3 minutes, 6 minutes, 4 minutes and 4 minutes. This is only an exemplary description and will be determined according to actual conditions.

[0082] Step S503: Determine at least one second difference value according to the filtered plurality of candidate second difference values.

[0083] Specifically, each unit time period included in the second historical time period corresponds to a second difference value, and the second difference value is an average value of multiple candidate second difference values ​​in each unit time period.

[0084] In this embodiment, assuming that the second historical period includes 3 unit periods, each unit period includes 6 orders, and the filtered multiple candidate second difference values ​​are 5 minutes, 4 minutes, 5 minutes, 5 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes, 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes and 4 minutes, which are divided into three groups. The first group is the candidate second difference values ​​corresponding to the first unit period, including 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes and 2 minutes; the second difference corresponding to the first unit period = (5+4+5+5+3 +2) minutes / 6=4 minutes; the second group is the candidate second difference values ​​corresponding to the second unit time period, including 6 minutes, 4 minutes, 4 minutes, 5 minutes, 3 minutes, and 5 minutes; the second difference value corresponding to the second unit time period = (6+4+4+5+3+5) minutes / 6 = 4.5 minutes; the third group is the candidate second difference values ​​corresponding to the third unit time period, including 4 minutes, 3 minutes, 3 minutes, 6 minutes, 4 minutes, and 4 minutes; the second difference value corresponding to the third unit time period = (4+3+3+6+4+4) minutes / 6 = 4 minutes; this is only an exemplary description and shall be determined according to actual conditions.

[0085] Step S504: Determine the sum of the at least one second difference value as the second difference signal.

[0086] The second difference corresponding to the first unit time period is 4 minutes, the second difference corresponding to the second unit time period is 4.5 minutes, and the second difference corresponding to the third unit time period is 4 minutes. The sum = 4 minutes + 4.5 minutes + 4 minutes = 12.5 minutes, that is, the second difference signal is 12.5 minutes.

[0087] In this embodiment, the specific process of obtaining the third difference signal is as follows: Figure 6 As shown, Figure 6 FIG. 1 is a flow chart of a method for obtaining a third difference signal according to the first embodiment of the present invention. Figure 6 As shown, the method of this embodiment includes the following steps:

[0088] Step S601, obtaining the historical prediction time difference of multiple orders in the first historical period, and the historical prediction time difference of multiple orders in the third historical period.

[0089] For example, obtain the historical predicted time difference TGAP of multiple orders within N hours before the corresponding moment of the current order, and the historical predicted time difference TGAP of multiple orders within N+1 hours before the corresponding moment of the current order, where N is a preset value, which can be an integer or a decimal, and is determined according to the user's needs. Assuming that the user needs to predict the delivery time at a fine granularity, N can be set to 1 hour, that is, with 1 hour as a cycle, obtain the historical predicted time difference TGAP of multiple orders within 1 hour before the current order occurs, and obtain the historical predicted time of multiple orders within 2 hours before the current order occurs. Difference TGAP; assuming that there are 10 orders within 1 hour before the current order, the historical prediction time difference TGAP of multiple orders within 1 hour before the current order is 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes and 20 minutes respectively; assuming that there are 14 orders within 2 hours before the current order, the historical prediction time difference TGAP of multiple orders within 2 hours before the current order is 6 minutes, 4 minutes, 6 minutes, 5 minutes, 4 minutes, 3 minutes, 7 minutes, 5 minutes, 5 minutes, 20 minutes, 3 minutes, 2 minutes, 2 minutes and 2 minutes respectively.

[0090] Step S602: Perform outlier filtering on the historical prediction time differences of multiple orders in the first historical period and the historical prediction time differences of multiple orders in the third historical period through a signal filter, and determine the filtered historical prediction time differences of multiple orders in the first historical period and the filtered historical prediction time differences of multiple orders in the third historical period.

[0091] Specifically, outliers in the historical prediction time differences of multiple orders in the first historical period are filtered. For example, among the 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, 4 minutes and 20 minutes, the 20 minutes is an outlier, and the 20 minutes are deleted to determine the filtered historical prediction time differences of multiple orders in the first historical period, namely 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes and 4 minutes; outliers in the historical prediction time differences of multiple orders in the first historical period are filtered. The values ​​are filtered, for example, among the 6 minutes, 4 minutes, 6 minutes, 5 minutes, 4 minutes, 3 minutes, 7 minutes, 5 minutes, 5 minutes, 20 minutes, 3 minutes, 2 minutes, 2 minutes and 2 minutes, the 20 minutes is an outlier, and the 20 minutes is deleted to determine the historical prediction time difference of multiple orders in the filtered third historical period, that is, 6 minutes, 4 minutes, 6 minutes, 5 minutes, 4 minutes, 3 minutes, 7 minutes, 5 minutes, 5 minutes, 3 minutes, 2 minutes, 2 minutes and 2 minutes. This is only an exemplary description and shall be determined according to actual conditions.

[0092] Step S603, determining a first average value of the historical prediction time differences of the plurality of orders in the first historical period after filtering, and a second average value of the historical prediction time differences of the plurality of orders in the third historical period after filtering.

[0093] In this embodiment, assuming that the historical prediction time difference of multiple orders in the first historical period after filtering is 5 minutes, 3 minutes, 5 minutes, 4 minutes, 3 minutes, 2 minutes, 6 minutes, 4 minutes, and 4 minutes, then the first average value is (5+3+5+4+3+2+6+4+4) minutes / 9=4 minutes, then the first average value is 4 minutes; the historical prediction time difference of multiple orders in the third historical period after filtering is 6 minutes, 4 minutes, 6 minutes, 5 minutes, 4 minutes, 3 minutes, 7 minutes, 5 minutes, and 5 minutes, then the second average value is (6+4+6+5+4+3+7+5+5) minutes / 9=5 minutes; this is only an exemplary description and shall be determined according to actual conditions.

[0094] Step S604: determine the difference between the first average value and the second average value as the third difference signal.

[0095] In this embodiment, the first average value 4 minutes - the second average value 5 minutes = -1, that is, the third difference signal is -1.

[0096] In an embodiment of the present invention, three filtered difference signals are determined through the above three embodiments, which are respectively a first difference signal, a second difference signal and a third difference signal. The first difference signal, the second difference signal and the third difference signal are updated at set intervals. For example, the data for calculating the first difference signal, the second difference signal and the third difference signal are updated every 0.5 hours to ensure the real-time performance of the first difference signal, the second difference signal and the third difference signal.

[0097] Step S303: Generate a target difference signal according to the first difference signal, the second difference signal and the third difference signal.

[0098] Specifically, the target difference signal is the predicted difference between the predicted delivery time and the actual delivery time of the current order.

[0099] In this embodiment, the first product of the first weight and the target difference signal, the second product of the second weight and the second difference signal, and the third product of the third weight and the third difference signal are determined; the sum of the first product, the second product and the third product is determined as the target difference signal. It can be understood that the process of determining the first product, the second product and the third product by the first weight, the second weight and the third weight is the process of PID adjustment by the PID controller, the first product is proportional P adjustment, it can also be said that the first difference affects the current deviation of the PID controller, it can also be called the time difference; the second product is integral I adjustment, it can also be said that the second difference reduces the oscillation of the PID controller, it can also be called the accumulation of time; the third product is differential D adjustment, it can also be said that the third difference determines the change trend of the PID controller, making the time difference change more sensitive, it can also be called the gap data with the target time.

[0100] In this embodiment, the weight values ​​of the first weight, the second weight and the third weight of the PID controller are set according to the scene, wherein the scene includes a set time period and a set area, wherein the set time period is the time period of the current order, and the set area is the delivery area to which the current order belongs; specifically, the set area may be a business district, a residential area, etc., and the set time period may be a lunch dining period, a dinner dining period, a non-meal time period, etc. The weight value of the PID controller may be set according to the lunch dining period of the business district, or may be set according to the non-meal time period of the residential area, and is determined specifically according to needs, thereby achieving local adjustment of the delivery time in the set area and the set time period.

[0101] Step S304: adjusting the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order.

[0102] In this embodiment, the target difference signal is superimposed on the initial prediction result of the current order to generate the predicted delivery time of the current order.

[0103] In this embodiment, since the target difference signal is the predicted difference between the predicted delivery time and the actual delivery time of the current order, the predicted delivery time of the current order can be accurately determined based on the target difference signal and the initial prediction result.

[0104] For example, assuming that the target difference signal is 7 minutes and the initial prediction result is 30 minutes, the predicted delivery time of the current order is 30+7=37 minutes; assuming that the target difference signal is -5 minutes and the initial prediction result is 30 minutes, the predicted delivery time of the current order is 30+(-5)=25 minutes; this is only an example description and the specific time will be determined based on actual conditions.

[0105] In this embodiment, the PID controller can be used to adjust the predicted delivery time in real time, reducing the local fluctuation range of the predicted delivery time prediction error, making the predicted delivery time more stable and accurate; through the above method, the user experience can be improved, the order placement rate can be increased, and the rider delivery efficiency can be improved, which can promote the development and operational efficiency of the logistics industry as a whole. Through the real-time regulation of the PID controller, the occurrence of abnormal predictions can be reduced and the stability of business operations can be improved.

[0106] Figure 7 FIG. 1 is a flow chart of a delivery time prediction method in an optional implementation of the first embodiment of the present invention. Figure 7 As shown, in an optional implementation of this embodiment, after step S304, the method further includes:

[0107] Step S305: Send the predicted delivery time to the user terminal.

[0108] In this embodiment, after the user terminal receives the predicted delivery time, it is displayed to the user in real time, so that the user can know a more accurate predicted delivery time, which can increase the probability of placing an order.

[0109] Optionally, after the user places an order, the rider delivers the order. Based on the completion of the order, the server corresponding to the user terminal determines the actual delivery time. The server sends the actual delivery time to the prediction module. The prediction module receives the actual delivery time sent by the server and uses it to predict the delivery time of the next cycle.

[0110] In this embodiment, when designing the PID controller, not only different adaptive parameters are provided according to different time periods and different scenarios, but also an abnormality judgment logic is designed to reduce the impact of over-adjustment.

[0111] In the embodiment of the present invention, the PID controller can use similar cities or historical values ​​for adjustment feedback in extreme scenarios and extreme cities, that is, when there is no historical data, to achieve a cold start of the PID controller.

[0112] In the embodiment of the present invention, a PID controller is used to generate a target difference signal, which is not affected by external factors such as weather and operation difficulty. It is also not aware of time and only needs to obtain historical time data. It has stronger generalization and robustness, better stability, and more stable control performance in time.

[0113] Figure 8 FIG. 2 is a schematic diagram of a delivery time prediction device according to a second embodiment of the present invention. Figure 8 As shown, the delivery time prediction device of this embodiment includes an acquisition unit 801, a generation unit 802 and an adjustment unit 803.

[0114] Among them, the acquisition unit 801 is used to obtain the initial prediction result of the current order; the acquisition unit 801 is also used to obtain a first difference signal, a second difference signal and a third difference signal, wherein the first difference signal represents a first difference between the historical prediction time difference of the first historical period and the target prediction time difference, the second difference signal represents the sum of at least one second difference between the historical prediction time difference of the second historical period and the target prediction time difference, and the third difference signal represents the difference between the historical prediction time difference of the first historical period and the historical prediction time difference of the third historical period, the first historical period is a unit period before the current moment, and the second historical period is a unit period before the current moment. The segment is a period composed of multiple unit time periods before the current moment, the third historical period is the unit time period before the first historical period, and the predicted time difference is the difference between the predicted delivery time and the actual delivery time in the corresponding time period; the generating unit 802 is used to perform PID adjustment according to the first difference signal, the second difference signal and the third difference signal to generate a target difference signal, wherein the target difference signal is the predicted difference between the predicted delivery time and the actual delivery time of the current order; the adjusting unit 803 is used to adjust the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order.

[0115] In an optional implementation, the acquiring unit is specifically configured to:

[0116] Acquire a candidate first difference signal, wherein the candidate first difference signal represents a plurality of candidate first differences between historical prediction time differences and target prediction time differences of a plurality of orders in a first historical period;

[0117] Perform outlier filtering on the plurality of candidate first difference values ​​by using a signal filter to determine a plurality of filtered candidate first difference values;

[0118] An average value of the plurality of first difference candidates after filtering is determined as the first difference signal.

[0119] In an optional implementation, the acquiring unit is specifically configured to:

[0120] Acquire a candidate second difference signal, wherein the candidate second difference signal represents a plurality of candidate second differences between historical prediction time differences and target prediction time differences of a plurality of orders in a second historical period;

[0121] Perform outlier filtering on the plurality of candidate second difference values ​​by using a signal filter to determine a plurality of filtered candidate second difference values;

[0122] Determine at least one second difference value according to the filtered multiple candidate second difference values, wherein each unit period included in the second historical period corresponds to a second difference value, and the second difference value is an average value of multiple candidate second difference values ​​in each unit period;

[0123] The sum of the at least one second difference value is determined as the second difference signal.

[0124] In an optional implementation, the acquiring unit is specifically configured to:

[0125] Obtaining historical prediction time differences of multiple orders in a first historical period and historical prediction time differences of multiple orders in a third historical period;

[0126] Performing outlier filtering on the historical prediction time differences of the plurality of orders in the first historical period and the historical prediction time differences of the plurality of orders in the third historical period through a signal filter, and determining the filtered historical prediction time differences of the plurality of orders in the first historical period and the filtered historical prediction time differences of the plurality of orders in the third historical period;

[0127] Determine a first average value of the historical prediction time differences of the plurality of orders in the first filtered historical period, and a second average value of the historical prediction time differences of the plurality of orders in the third filtered historical period;

[0128] A difference between the first average value and the second average value is determined as the third difference signal.

[0129] In an optional implementation, the generating unit is specifically used for:

[0130] Determine a first product of a first weight and the target difference signal, a second product of a second weight and the second difference signal, and a third product of a third weight and the third difference signal; and determine a sum of the first product, the second product and the third product as the target difference signal.

[0131] In an optional implementation, the acquiring unit is specifically configured to:

[0132] Get the order characteristics of the current order;

[0133] The order features are input into a pre-trained machine learning model to determine an initial prediction result for the current order.

[0134] In an optional implementation, the adjusting unit is specifically configured to:

[0135] The target difference signal is superimposed on the initial prediction result of the current order to generate a predicted delivery time for the current order.

[0136] In an optional implementation, the device further includes:

[0137] The sending unit is used to send the predicted delivery time to the user terminal.

[0138] In an optional implementation, the first difference signal represents a first difference between a historical prediction time difference in a set area for a first historical period and a target prediction time difference, the second difference signal represents a sum of at least one second difference between a historical prediction time difference in a set area for a second historical period and a target prediction time difference, and the third difference signal represents a difference between a historical prediction time difference in the first historical period in the set area and a historical prediction time difference in the third historical period in the set area, and the set area is the delivery area to which the current order belongs.

[0139] In an optional implementation, the first weight, the second weight and the third weight are set according to a scenario, wherein the scenario includes a set time period and a set area, wherein the set time period is a unit time period in which the current order is located, and the set area is a delivery area to which the current order belongs.

[0140] In an optional implementation, the device further includes:

[0141] An updating unit is used to update the first difference signal, the second difference signal and the third difference signal at every set period.

[0142] In the embodiment of the present invention, the initial prediction result of the current order is obtained; the first difference signal, the second difference signal and the third difference signal are obtained, and a target difference signal is generated according to the first difference signal, the second difference signal and the third difference signal, wherein the target difference signal is the predicted difference between the predicted delivery time and the actual delivery time of the current order; the initial prediction result of the current order is adjusted according to the target difference signal to determine the predicted delivery time of the current order. Therefore, in the embodiment of the present invention, the above method can be used to make real-time predictions on the delivery time of orders in a local area, and the delivery time of orders can be predicted quickly and accurately.

[0143] Fig. 9 Schematic diagram of an electronic device according to a third embodiment of the present invention. In this embodiment, the electronic device 9 includes a server, a terminal, etc. Fig. 9 As shown, the electronic device 9 includes: at least one processor 91; a memory 92 connected to the at least one processor 91 for communication; and a communication component 93 connected to the scanning device for communication, the communication component 93 receives and sends data under the control of the processor 91; wherein the memory 92 stores instructions executable by at least one processor 91, and the instructions are executed by at least one processor 91 to implement the above-mentioned delivery time prediction method.

[0144] Specifically, the electronic device includes: one or more processors 91 and a memory 92, Fig. 9 A processor 91 is taken as an example. The processor 91 and the memory 92 may be connected via a bus or other means. Fig. 9 In the example, the bus connection is used. The memory 92 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The processor 91 executes various functional applications and data processing of the device by running the non-volatile software programs, instructions and modules stored in the memory 92, that is, realizing the above-mentioned delivery time prediction method.

[0145] The memory 92 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and applications required for at least one function; the data storage area may store a list of options, etc. In addition, the memory 92 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 92 may optionally include a memory remotely arranged relative to the processor 91, and these remote memories may be connected to an external device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0146] One or more modules are stored in the memory 92, and when executed by one or more processors 91, the delivery time prediction method in any of the above method embodiments is executed.

[0147] The above-mentioned product can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present application.

[0148] The embodiment of the present invention obtains the initial prediction result of the current order; obtains the first difference signal, the second difference signal and the third difference signal, performs PID adjustment on the first difference signal, the second difference signal and the third difference signal to generate a target difference signal, wherein the target difference signal is the predicted difference between the predicted delivery time and the actual delivery time of the current order; adjusts the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order. Therefore, in the embodiment of the present invention, the PID adjustment can be used to make real-time predictions on the delivery time of orders in a local area, and the delivery time of the orders can be predicted quickly and accurately.

[0149] Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program, wherein the computer-readable program is used for a computer to execute part or all of the above method embodiments.

[0150] That is, those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including a number of instructions to enable a device (which can be a single-chip microcomputer, chip, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0151] The embodiment of the present invention discloses A1, a delivery time prediction method, the method comprising:

[0152] Get the initial forecast result of the current order;

[0153] Acquire a first difference signal, a second difference signal, and a third difference signal, wherein the first difference signal represents a first difference between a historical prediction time difference and a target prediction time difference in a first historical period, the second difference signal represents a sum of at least one second difference between a historical prediction time difference and a target prediction time difference in a second historical period, and the third difference signal represents a difference between a historical prediction time difference in the first historical period and a historical prediction time difference in a third historical period, the first historical period is a unit period before the current moment, the second historical period is a period composed of a plurality of unit periods before the current moment, the third historical period is a unit period before the first historical period, and the prediction time difference is a difference between a predicted delivery time and an actual delivery time in a corresponding time period;

[0154] generating a target difference signal according to the first difference signal, the second difference signal and the third difference signal, wherein the target difference signal is a predicted difference between a predicted delivery time and an actual delivery time of the current order;

[0155] The initial prediction result of the current order is adjusted according to the target difference signal to determine the predicted delivery time of the current order.

[0156] A2. According to the method described in A1, obtaining the first difference signal, the second difference signal and the third difference signal comprises:

[0157] Acquire a candidate first difference signal, wherein the candidate first difference signal represents a plurality of candidate first differences between historical prediction time differences and target prediction time differences of a plurality of orders in a first historical period;

[0158] Perform outlier filtering on the plurality of candidate first difference values ​​by using a signal filter to determine a plurality of filtered candidate first difference values;

[0159] An average value of the plurality of first difference candidates after filtering is determined as the first difference signal.

[0160] A3. According to the method described in A1, the obtaining of the first difference signal, the second difference signal and the third difference signal comprises:

[0161] Acquire a candidate second difference signal, wherein the candidate second difference signal represents a plurality of candidate second differences between historical prediction time differences and target prediction time differences of a plurality of orders in a second historical period;

[0162] Perform outlier filtering on the plurality of candidate second difference values ​​by using a signal filter to determine a plurality of filtered candidate second difference values;

[0163] Determine at least one second difference value according to the filtered multiple candidate second difference values, wherein each unit period included in the second historical period corresponds to a second difference value, and the second difference value is an average value of multiple candidate second difference values ​​in each unit period;

[0164] The sum of the at least one second difference value is determined as the second difference signal.

[0165] A4. According to the method described in A1, obtaining the first difference signal, the second difference signal and the third difference signal comprises:

[0166] Obtaining historical prediction time differences of multiple orders in a first historical period and historical prediction time differences of multiple orders in a third historical period;

[0167] Performing outlier filtering on the historical prediction time differences of the plurality of orders in the first historical period and the historical prediction time differences of the plurality of orders in the third historical period through a signal filter, and determining the filtered historical prediction time differences of the plurality of orders in the first historical period and the filtered historical prediction time differences of the plurality of orders in the third historical period;

[0168] Determine a first average value of the historical prediction time differences of the plurality of orders in the first filtered historical period, and a second average value of the historical prediction time differences of the plurality of orders in the third filtered historical period;

[0169] A difference between the first average value and the second average value is determined as the third difference signal.

[0170] A5. The method according to A1, wherein generating a target difference signal according to the first difference signal, the second difference signal and the third difference signal comprises:

[0171] determining a first product of a first weight and the target difference signal, a second product of a second weight and the second difference signal, and a third product of a third weight and the third difference signal;

[0172] A sum of the first product, the second product and the third product is determined as the target difference signal.

[0173] A6. According to the method described in A1, obtaining the initial prediction result of the current order specifically includes:

[0174] Get the order characteristics of the current order;

[0175] The order features are input into a pre-trained machine learning model to determine an initial prediction result for the current order.

[0176] A7. According to the method of A1, adjusting the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order specifically includes:

[0177] The target difference signal is superimposed on the initial prediction result of the current order to generate a predicted delivery time for the current order.

[0178] A8. The method according to A1, further comprising:

[0179] The predicted delivery time is sent to the user terminal.

[0180] A9. According to the method described in A1, the first difference signal represents a first difference between a historical prediction time difference in a set area for a first historical period and a target prediction time difference, the second difference signal represents a sum of at least one second difference between a historical prediction time difference in a set area for a second historical period and a target prediction time difference, and the third difference signal represents a difference between a historical prediction time difference in the first historical period within the set area and a historical prediction time difference in the third historical period within the set area, and the set area is the delivery area to which the current order belongs.

[0181] A10. According to the method described in A5, the first weight, the second weight and the third weight are set according to a scenario, wherein the scenario includes a set time period and a set area, wherein the set time period is a unit time period in which the current order is located, and the set area is a delivery area to which the current order belongs.

[0182] A11. The method according to A1, further comprising:

[0183] The first difference signal, the second difference signal and the third difference signal are updated every set period.

[0184] The embodiment of the present invention further discloses B1, a delivery time prediction device, the device comprising:

[0185] An acquisition unit, used to acquire the initial prediction result of the current order;

[0186] The acquisition unit is further used to acquire a first difference signal, a second difference signal and a third difference signal, wherein the first difference signal represents a first difference between a historical prediction time difference and a target prediction time difference in a first historical period, the second difference signal represents a sum of at least one second difference between a historical prediction time difference and a target prediction time difference in a second historical period, and the third difference signal represents a difference between a historical prediction time difference in the first historical period and a historical prediction time difference in a third historical period, the first historical period is a unit period before the current moment, the second historical period is a period composed of a plurality of unit periods before the current moment, the third historical period is a unit period before the first historical period, and the prediction time difference is a difference between a predicted delivery time and an actual delivery time in a corresponding time period;

[0187] a generating unit, configured to perform PID adjustment according to the first difference signal, the second difference signal, and the third difference signal to generate a target difference signal, wherein the target difference signal is a predicted difference between a predicted delivery time and an actual delivery time of the current order;

[0188] An adjustment unit is used to adjust the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order.

[0189] An embodiment of the present invention also discloses C1, an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement a method as described in any one of A1-A11.

[0190] The embodiment of the present invention further discloses D1, a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in any one of A1-A11 is implemented.

[0191] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A delivery time prediction method, characterized in that: The method comprises: Get the initial forecast result of the current order; Acquire a first difference signal, a second difference signal, and a third difference signal, wherein the first difference signal represents a first difference between a historical prediction time difference and a target prediction time difference in a first historical period, the second difference signal represents a sum of at least one second difference between a historical prediction time difference and a target prediction time difference in a second historical period, and the third difference signal represents a difference between a historical prediction time difference in the first historical period and a historical prediction time difference in a third historical period, the first historical period is a unit period before the current moment, the second historical period is a period composed of a plurality of unit periods before the current moment, the third historical period is a unit period before the first historical period, and the prediction time difference is a difference between a predicted delivery time and an actual delivery time in a corresponding time period; generating a target difference signal according to the first difference signal, the second difference signal and the third difference signal, wherein the target difference signal is a predicted difference between a predicted delivery time and an actual delivery time of the current order; The initial prediction result of the current order is adjusted according to the target difference signal to determine the predicted delivery time of the current order.

2. The method according to claim 1, characterized in that The obtaining of the first difference signal, the second difference signal and the third difference signal comprises: Acquire a candidate first difference signal, wherein the candidate first difference signal represents a plurality of candidate first differences between historical prediction time differences and target prediction time differences of a plurality of orders in a first historical period; Perform outlier filtering on the plurality of candidate first difference values ​​by using a signal filter to determine a plurality of filtered candidate first difference values; An average value of the plurality of first difference candidates after filtering is determined as the first difference signal.

3. The method according to claim 1, characterized in that The obtaining of the first difference signal, the second difference signal and the third difference signal comprises: Acquire a candidate second difference signal, wherein the candidate second difference signal represents a plurality of candidate second differences between historical prediction time differences and target prediction time differences of a plurality of orders in a second historical period; Perform outlier filtering on the plurality of candidate second difference values ​​by using a signal filter to determine a plurality of filtered candidate second difference values; Determine at least one second difference value according to the filtered multiple candidate second difference values, wherein each unit period included in the second historical period corresponds to a second difference value, and the second difference value is an average value of multiple candidate second difference values ​​in each unit period; The sum of the at least one second difference value is determined as the second difference signal.

4. The method according to claim 1, characterized in that: The obtaining of the first difference signal, the second difference signal and the third difference signal comprises: Obtaining historical prediction time differences of multiple orders in a first historical period and historical prediction time differences of multiple orders in a third historical period; Performing outlier filtering on the historical prediction time differences of the plurality of orders in the first historical period and the historical prediction time differences of the plurality of orders in the third historical period through a signal filter, and determining the filtered historical prediction time differences of the plurality of orders in the first historical period and the filtered historical prediction time differences of the plurality of orders in the third historical period; Determine a first average value of the historical prediction time differences of the plurality of orders in the first filtered historical period, and a second average value of the historical prediction time differences of the plurality of orders in the third filtered historical period; A difference between the first average value and the second average value is determined as the third difference signal.

5. The method according to claim 1, characterized in that The generating a target difference signal according to the first difference signal, the second difference signal and the third difference signal specifically includes: determining a first product of a first weight and the target difference signal, a second product of a second weight and the second difference signal, and a third product of a third weight and the third difference signal; A sum of the first product, the second product and the third product is determined as the target difference signal.

6. The method according to claim 1, characterized in that The obtaining of the initial prediction result of the current order specifically includes: Get the order characteristics of the current order; The order features are input into a pre-trained machine learning model to determine an initial prediction result for the current order.

7. The method according to claim 1, characterized in that The step of adjusting the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order specifically includes: The target difference signal is superimposed on the initial prediction result of the current order to generate a predicted delivery time for the current order.

8. A delivery time prediction device, characterized in that: The device comprises: An acquisition unit, used to acquire the initial prediction result of the current order; The acquisition unit is further used to acquire a first difference signal, a second difference signal and a third difference signal, wherein the first difference signal represents a first difference between a historical prediction time difference and a target prediction time difference in a first historical period, the second difference signal represents a sum of at least one second difference between a historical prediction time difference and a target prediction time difference in a second historical period, and the third difference signal represents a difference between a historical prediction time difference in the first historical period and a historical prediction time difference in a third historical period, the first historical period is a unit period before the current moment, the second historical period is a period composed of a plurality of unit periods before the current moment, the third historical period is a unit period before the first historical period, and the prediction time difference is a difference between a predicted delivery time and an actual delivery time in a corresponding time period; a generating unit, configured to perform PID adjustment according to the first difference signal, the second difference signal, and the third difference signal to generate a target difference signal, wherein the target difference signal is a predicted difference between a predicted delivery time and an actual delivery time of the current order; An adjustment unit is used to adjust the initial prediction result of the current order according to the target difference signal to determine the predicted delivery time of the current order.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.