CN110909908A - Method and device for predicting item picking duration - Google Patents

Method and device for predicting item picking duration Download PDF

Info

Publication number
CN110909908A
CN110909908A CN201811087230.0A CN201811087230A CN110909908A CN 110909908 A CN110909908 A CN 110909908A CN 201811087230 A CN201811087230 A CN 201811087230A CN 110909908 A CN110909908 A CN 110909908A
Authority
CN
China
Prior art keywords
picking
determining
kernel function
characteristic data
sample object
Prior art date
Legal status (The legal status 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 status listed.)
Granted
Application number
CN201811087230.0A
Other languages
Chinese (zh)
Other versions
CN110909908B (en
Inventor
范超
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Jingdong Qianshi Technology Co Ltd
Original Assignee
Tianjin Jingdong Shentuo Robot Technology Co ltd
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 Tianjin Jingdong Shentuo Robot Technology Co ltd filed Critical Tianjin Jingdong Shentuo Robot Technology Co ltd
Priority to CN201811087230.0A priority Critical patent/CN110909908B/en
Publication of CN110909908A publication Critical patent/CN110909908A/en
Application granted granted Critical
Publication of CN110909908B publication Critical patent/CN110909908B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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"
    • 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
    • G06Q10/087—Inventory or stock management, e.g. order filling, procurement or balancing against orders
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30—Computing systems specially adapted for manufacturing

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Economics (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Development Economics (AREA)
  • Theoretical Computer Science (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Game Theory and Decision Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a method and a device for predicting item picking time, and relates to the field of logistics storage. One embodiment of the method comprises: acquiring characteristic data and historical picking duration of a sample object; determining a kernel matrix based on a preset kernel function and the characteristic data of the sample object; and determining a picking duration prediction model according to the historical picking durations of the sample objects and the kernel matrix. The method overcomes the problem that the characteristic information influencing the object picking time is seriously insufficient, determines the picking time prediction model, improves the prediction accuracy, reduces the error rate and reduces the workload in logistics simulation.

Description

Method and device for predicting item picking duration
Technical Field
The invention relates to the field of logistics storage, in particular to a method and a device for predicting item picking time.
Background
The picking operation is an important component of the warehousing operation, wherein the time efficiency of the picking operation directly influences the time efficiency of the ex-warehouse and in-warehouse operation. Therefore, in the simulation of logistics, the picking time length needs to be predicted according to the related information of the goods, namely, the time period from when the picker sees the screen picking commander to when the picker picks the last goods of the same kind and then scans the goods grid needs to be known in the picking process of the goods.
For the logistics simulation, the picking time length of each commodity needs to be determined, and then the picking time length and the corresponding commodity are input into the logistics simulation system. In the prior art, the picking time length of a commodity is determined manually according to self experience, or historical picking data of the commodity is collected, and the picking time length of the commodity is determined according to the historical picking data. Because in the logistics simulation, the required data is more, if the picking time of each commodity under various conditions is determined manually, the workload is very large, the accuracy is low, the error rate is high, and the like.
And for the sorting duration of the commodity, the extractable characteristic data is less, so that the problem of under-fitting caused by insufficient characteristics is likely to be caused when the prediction model is built, namely, the information quantity possessed by the commodity is difficult to build a proper model. Moreover, most of the existing models need to assume a model form in advance, but the functional relationship between the commodity picking time and the characteristic data is very complex, so that the model form is difficult to assume accurately.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method and an apparatus for predicting an item picking duration, which overcome the problem of serious shortage of characteristic information affecting an object picking duration, and determine a picking duration prediction model, thereby improving the prediction accuracy, reducing the error rate, and reducing the workload in logistics simulation.
To achieve the above object, according to one aspect of the embodiments of the present invention, there is provided a method for item picking duration prediction.
The method for predicting the item picking time length comprises the following steps:
acquiring characteristic data and historical picking duration of a sample object;
determining a kernel matrix based on a preset kernel function and the characteristic data of the sample object;
and determining a picking duration prediction model according to the historical picking durations of the sample objects and the kernel matrix.
Optionally, the step of determining a kernel matrix based on a preset kernel function and the feature data of the sample object includes:
mapping the characteristic data of the sample object to an infinite dimension based on a preset kernel function; the preset kernel function is a Gaussian kernel function;
and determining a kernel matrix according to the mapping result.
Optionally, the step of mapping the feature data of the sample object to an infinite dimension based on a preset kernel function includes:
determining regulation and control parameters of a preset kernel function through ten-fold cross validation;
and mapping the characteristic data of the sample object to infinite dimensionality according to the regulation and control parameter and a preset kernel function.
Optionally, the step of determining a picking duration prediction model according to the historical picking durations of the sample objects and the kernel matrix includes:
the model parameters α are determined by the following formula:
Figure BDA0001803457210000031
wherein, K is a nuclear matrix,
Figure BDA0001803457210000032
selecting the historical picking time length of the sample object;
based on the model parameters α, determining a picking time duration prediction model as:
y＝[1+k(x1,x),…,1+k(xn,x)]α
where y is the picking duration of the object, x1,...,xNCharacteristic data of the n sample objects respectively; k (x)iX) is a kernel function, i ═ 1, 2.
Optionally, after establishing a picking duration prediction model according to the historical picking durations of the sample objects and the kernel matrix, the method further includes:
reducing the number of the feature data according to a preset step length, and taking the reduced feature data as a feature data set to be calculated;
determining the mean square error of each feature data set to be calculated based on the determined sorting duration prediction model;
taking the characteristic data set to be calculated corresponding to the minimum mean square error as an adjustment characteristic data set;
adjusting the picking length prediction model based on the adjusted feature data set.
Optionally, the characteristic data of the sample object at least comprises: number of single picks, weight, volume, number of shelves on which the sample object is located.
To achieve the above object, according to another aspect of the embodiments of the present invention, there is provided an apparatus for item picking duration prediction.
The device for predicting the item picking time length of the embodiment of the invention comprises the following steps:
the sample data acquisition module is used for acquiring the characteristic data and the historical picking duration of the sample object;
the kernel matrix determining module is used for determining a kernel matrix based on a preset kernel function and the characteristic data of the sample object;
and the model determining module is used for determining a picking duration prediction model according to the historical picking duration of the sample object and the nuclear matrix.
Optionally, the kernel matrix determination module is further configured to map feature data of the sample object to an infinite dimension based on a preset kernel function; the preset kernel function is a Gaussian kernel function; and determining a kernel matrix according to the mapping result.
Optionally, the kernel matrix determination module is further configured to determine a preset regulation parameter of the kernel function through cross validation by ten folds; and mapping the characteristic data of the sample object to infinite dimensionality according to the regulation and control parameter and a preset kernel function.
Optionally, the model determination module is further configured to determine model parameters α by:
Figure BDA0001803457210000041
wherein, K is a nuclear matrix,
Figure BDA0001803457210000042
selecting the historical picking time length of the sample object;
based on the model parameters α, determining a picking time duration prediction model as:
y＝[1+k(x1,x),...,1+k(xn,x)]α
where y is the picking duration of the object, x1,...,xNCharacteristic data of the n sample objects respectively; k (x)iX) is a kernel function, i ═ 1, 2.
Optionally, the system further comprises an adjusting module, configured to reduce the number of the feature data according to a preset step length, and use the reduced feature data as a feature data set to be calculated; determining the mean square error of each feature data set to be calculated based on the determined sorting duration prediction model; taking the characteristic data set to be calculated corresponding to the minimum mean square error as an adjustment characteristic data set; adjusting the picking length prediction model based on the adjusted feature data set.
To achieve the above object, according to still another aspect of an embodiment of the present invention, there is provided an electronic apparatus.
The electronic device of the embodiment of the invention comprises: one or more processors; storage means for storing one or more programs which, when executed by the one or more processors, cause the one or more processors to implement any of the above-described methods for item picking time prediction.
To achieve the above object, according to a further aspect of the embodiments of the present invention, there is provided a computer-readable medium having stored thereon a computer program, wherein the program, when executed by a processor, implements any of the above-mentioned methods for item picking duration prediction.
One embodiment of the above invention has the following advantages or benefits: since the kernel function can expand information of finite dimensions into information of infinite dimensions, the technical scheme solves the problem of insufficient characteristic information by using the kernel function. And based on the kernel function, the functional relation between the picking time and the characteristic data can be conveniently determined, and the problem that the existing functional model needs to assume a model form in advance is solved. The embodiment of the invention not only improves the accuracy of prediction and reduces the error rate, but also reduces the workload in logistics simulation.
Further effects of the above-mentioned non-conventional alternatives will be described below in connection with the embodiments.
Drawings
The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:
FIG. 1 is a schematic illustration of a main flow of a method for item picking duration prediction according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a method for item picking duration prediction according to an embodiment of the present invention;
FIG. 3 is a schematic diagram of the main modules of an apparatus for item picking duration prediction according to an embodiment of the present invention;
FIG. 4 is an exemplary system architecture diagram in which embodiments of the present invention may be employed;
fig. 5 is a schematic block diagram of a computer system suitable for use in implementing a terminal device or server of an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present invention are described below with reference to the accompanying drawings, in which various details of embodiments of the invention are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
Fig. 1 is a schematic diagram of a main flow of a method for item picking duration prediction according to an embodiment of the present invention, and as shown in fig. 1, the method for item picking duration prediction according to an embodiment of the present invention mainly includes:
step S101: characteristic data and historical picking duration of the sample object are obtained. The characteristic data of the sample object at least comprises: number of single picks, weight, volume, number of shelves on which the sample object is located. The number of single picking pieces is the number of picking objects in each picking operation; the weight refers to the total weight of the object in the picking; volume refers to the total volume of the subject in the picking; the number of shelf levels at which the sample object is located indicates the location of the shelf at which the object is located. In order to improve the accuracy of the determined model, the picking time length of the object is the real picking time length in the historical picking data, namely the picking personnel sees the screen picking command time point y1 and the difference y of the time point y2 of scanning the goods grids after the picking personnel finishes picking the last one of the same type of goods, and the picking time length y is y2-y 1. In another embodiment of the invention, the picking duration of the sample object may also be determined manually based on experience, and is not necessarily the actual picking duration data.
Step S102: and determining a kernel matrix based on the preset kernel function and the characteristic data of the sample object. The kernel Function includes a linear kernel Function, a polynomial kernel Function, a gaussian kernel Function, etc., where the gaussian kernel Function is most commonly used, and can map data to an infinite dimension, which is also called a Radial Basis Function (RBF for short), and is a certain scalar Function symmetrical along a Radial direction. Generally defined as a monotonic function of the euclidean distance between any point x in space and some center xc, which can be written as k (| | x-xc |). The original existing characteristic data of the object can be mapped to a high-dimensional (even infinite-dimensional) regeneration kernel Hilbert space from an Euclidean space through a kernel function, namely xi→Φ(xi)，φ(xi) Is xiAnd mapping the feature vector.
Specifically, based on a preset kernel function, mapping feature data of a sample object to infinite dimensionality; the preset kernel function is a gaussian kernel function. And determining a kernel matrix according to the mapping result. Determining a preset regulation parameter of the kernel function through cross validation by ten folds; and mapping the characteristic data of the sample object to infinite dimensionality according to the regulation and control parameter and a preset kernel function. The most commonly used Radial Basis Function is a gaussian kernel Function, a so-called Radial Basis Function (RBF), which is a scalar Function that is radially symmetric. The Gaussian kernel function directly maps variables of an input space into an infinite dimensional space, can effectively solve the problem of insufficient information quantity of the input space, and does not need to know a specific mapping function relationship. The Gaussian kernel function is:
Figure BDA0001803457210000071
wherein, σ is a width parameter (regulation parameter) of the function, and controls the radial action range of the function. The taylor expansion is brought into the gaussian kernel to obtain an infinite dimension mapping:
Figure BDA0001803457210000072
then, for x1And x2The inner product form of (a) corresponds to the inner product calculation at infinite dimension in SVM, i.e. the gaussian kernel maps the data to a dimension of infinite height.
And ten-fold cross validation, called 10-fold cross-validation by English name, for testing the accuracy of the algorithm. Is a commonly used test method. The data set was divided into ten parts, and 9 parts of the data set were used as training data and 1 part of the data set was used as test data in turn for the experiments. Each trial will yield a corresponding accuracy (or error rate). The average of the accuracy (or error rate) of the 10 results is used as an estimate of the accuracy of the algorithm, and generally 10-fold cross validation is performed multiple times (for example, 10 times of 10-fold cross validation), and then the average is obtained as an estimate of the accuracy of the algorithm.
Step S103, determining a picking duration prediction model according to the historical picking duration and the kernel matrix of the sample object, specifically, determining model parameters α according to the following formula:
Figure BDA0001803457210000081
wherein, K is a nuclear matrix,
Figure BDA0001803457210000082
the historical picking durations for the sample objects.
Establishing a sorting duration model, known easily, of
Figure BDA0001803457210000083
Wherein,
(X′X)-2＝(X′X)-1(X′X)-1，α＝X(X′X)-2X′y
consider the mapping phi, order
Figure BDA0001803457210000084
Wherein x is1，x2，……，xnN sample points in the sample object feature data set X.
And the number of the first and second groups,
Figure BDA0001803457210000085
Figure BDA0001803457210000086
X(X′X)X′α＝XX′y
thus, correspondingly also have
Figure BDA0001803457210000087
Order core matrix
Figure BDA0001803457210000088
Then there are
KKα＝Ky
Thus, there are
α＝K-1y
Based on the model parameters α, determining a picking time duration prediction model as:
y＝[1+k(x1,x),…,1+k(xn,x)]α
where y is the picking duration of the object, x1,...,xNCharacteristic data of the n sample objects respectively; k (x)iX) is a kernel function, i ═ 1, 2.
From the above, the embodiment of the invention has a definite solving expression based on the least square method and the quadratic loss function. Compared with the support vector machine which is a change loss function, the method has the advantages that no clear solving expression exists, the calculation speed is low, variable selection is not easy to realize, and the calculation speed is high.
After the above process, the number of the feature data is reduced according to the preset step length, and the reduced feature data is used as a feature data set to be calculated. All variables may be stepped down by 1 variable at a time until the reduction is 1 variable, and the mean square error over the test set is calculated each time (the square of the difference between the actual and predicted values for each sample point is squared and then averaged). Therefore, based on the determined picking duration prediction model, the mean square error of each feature data set to be calculated is determined. And taking the characteristic data set to be calculated corresponding to the minimum mean square error as an adjustment characteristic data set, and adjusting the sorting duration prediction model based on the adjustment characteristic data set. Since the gaussian kernel is selected, as described above, to map a few variables in the original input space to an infinite number of variables in the hilbert space, the under-fitting problem may become an over-fitting problem.
Assuming a sample size of 100, there are 4 original independent variables (weight, volume, number of pieces, number of layers) and the dependent variable is the picking time. First, 70 sample points are randomly selected as a training set, and 30 sample points are selected as a testing set. Before calculation, all sample data is divided into a training set and a test set, the latter typically accounting for 30% of the total sample size. Then removing a variable, such as removing a volume, and calculating its corresponding mse1 (mean square error); the volume is retained, the weight is removed, and its corresponding mse2 is calculated, and so on, so that there can be 4 mean square errors mse. The minimum mse is found, assumed to be mse 1. And then, one is removed from the weight, number of pieces, number of layers, in a manner similar to the foregoing. Until only one variable remains. Then comparing which mse is smaller when 1 variable, 2 variables and 3 variables are removed, wherein the corresponding variable set is the last variable, and finally substituting into a formula for calculation to adjust the model. For example, after the above calculation, it is determined that the mean square error of the feature data set "weight, number of pieces" is the smallest in all the feature data combinations, the adjusted model is determined to relate to both the weight and the number of pieces, when the picking time period of the commodity is predicted based on the model, the weight and the number of pieces of the commodity are directly acquired (other feature data related in the training process are not acquired), and the picking time period of the commodity is calculated based on the adjusted model.
The above method for predicting the item picking duration is further described by taking an unmanned bin which takes an automatic guided vehicle AGV as a main robot as an example. FIG. 2 is a schematic diagram of a method for item picking duration prediction according to an embodiment of the present invention.
For the embodiment of the invention, the kernel function can expand the information with limited dimension into the information with infinite dimension, so the technical scheme solves the problem of insufficient characteristic information by using the kernel function. And based on the kernel function, the functional relation between the picking time and the characteristic data can be conveniently determined, and the problem that the existing functional model needs to assume a model form in advance is solved. The embodiment of the invention not only improves the accuracy of prediction and reduces the error rate, but also reduces the workload in logistics simulation. Also, to prevent overfitting, embodiments of the present invention are controlled by way of variable selection.
As shown in fig. 2, a method for item picking duration prediction of an embodiment of the present invention includes:
step S201: characteristic data of the sample commodity is obtained. In the unmanned storehouse, commodities are stored on shelves, each shelf is provided with a plurality of layers, and each layer is divided into different goods grids. After the user places an order, the AGV trolley receives a corresponding instruction, firstly finds a proper goods shelf and then carries to the workstation. The order picker will then pick a certain number of items from a certain compartment on a certain level of the shelf according to the instructions on the computer screen. Due to the availability of data, only the number of single pickers, the weight, the volume and the number of shelves on which the goods are placed in the same type of goods can be extracted, and the characteristic data matrix is marked as X.
Step S202: historical picking durations for sample items are collected. Recording the time point when the picking personnel sees the screen picking command, setting the time point as y1, recording the time of scanning the goods grid after the picking personnel picks the last one of the same type of goods as y2, and taking y as y2-y1 as the historical picking time length y of the sample goods.
Step S203: and mapping the characteristic data of the original existing commodity from an Euclidean space to a regeneration kernel Hilbert space with infinite dimensions through a Gaussian kernel function. Hilbert space refers to the complete inner product space. It can be seen as a generalization of the european space. In Euclidean space, the vectors are finite dimensional and define inner products. If the inner product space of finite dimension is expanded to the inner product space of infinite dimension, the space is Hilbert space. Whether a space is complete is determined by convergence of all Cauchy (Cauchy) columns in the space.
Step S204: calculating a kernel matrix K:
Figure BDA0001803457210000111
and, an unknown parameter σ in the kernel function2Determined by means of 10-fold cross validation.
Step S205, determining model parameters α according to the historical picking duration of the sample object and the kernel matrix K:
Figure BDA0001803457210000112
step S206: the model is adjusted by a technique of backward variable selection. And substituting all independent variables in the original input space into the calculation, and calculating the mean square error on the test set. Each time, 1 variable is decremented, all variables are tried step by step until the decrement is 1 variable, the mean square error over the test set is calculated each time. And selecting the variable set corresponding to the minimum mean square error as the finally selected variable set.
The embodiment of the invention solves the problem in prediction in a kernel function mode. The kernel function may expand information of a finite dimension to information of an infinite dimension. Since the data is mapped to a high-dimensional Regenerative Kernel Hilbert Space (RKHS), there may be an overfitting problem, and in order to prevent overfitting, the embodiment of the present invention is controlled by way of variable selection.
Fig. 3 is a schematic diagram of main modules of an apparatus for item picking duration prediction according to an embodiment of the present invention, and as shown in fig. 3, an apparatus 300 for item picking duration prediction according to an embodiment of the present invention includes a sample data acquisition module 301, a kernel matrix determination module 302, and a model determination module 303.
The sample data obtaining module 301 is configured to obtain feature data and historical picking duration of a sample object.
The kernel matrix determination module 302 is configured to determine a kernel matrix based on a preset kernel function and feature data of the sample object. The kernel matrix determination module is further used for mapping the characteristic data of the sample object to infinite dimensionality based on a preset kernel function; the preset kernel function is a Gaussian kernel function; and determining a kernel matrix according to the mapping result. The kernel matrix determining module is further used for determining regulation and control parameters of the preset kernel function through cross validation by ten folds; and mapping the characteristic data of the sample object to infinite dimensionality according to the regulation and control parameter and a preset kernel function.
The model determination module 303 is further operable to determine model parameters α by:
Figure BDA0001803457210000121
whereinAnd K is a kernel matrix,
Figure BDA0001803457210000122
selecting the historical picking time length of the sample object;
based on model parameters α, the picking duration prediction model is determined as:
y＝[1+k(x1,x),...,1+k(xn,x)]α
where y is the picking duration of the object, x1,...,xNCharacteristic data of the n sample objects respectively; k (x)iX) is a kernel function, i ═ 1, 2.
The device for predicting the item picking duration of the embodiment of the invention further comprises an adjusting module, a calculating module and a calculating module, wherein the adjusting module is used for reducing the number of the characteristic data according to the preset step length and taking the reduced characteristic data as a characteristic data set to be calculated; determining the mean square error of each feature data set to be calculated based on the determined sorting duration prediction model; taking the characteristic data set to be calculated corresponding to the minimum mean square error as an adjustment characteristic data set; the picking duration prediction model is adjusted based on the adjusted feature data set.
For the embodiment of the invention, the kernel function can expand the information with limited dimension into the information with infinite dimension, so the technical scheme solves the problem of insufficient characteristic information by using the kernel function. And based on the kernel function, the functional relation between the picking time and the characteristic data can be conveniently determined, and the problem that the existing functional model needs to assume a model form in advance is solved. The embodiment of the invention not only improves the accuracy of prediction and reduces the error rate, but also reduces the workload in logistics simulation. Also, to prevent overfitting, embodiments of the present invention are controlled by way of variable selection.
Fig. 4 illustrates an exemplary system architecture 400 for a method for item pick time prediction or an apparatus for item pick time prediction to which embodiments of the present invention may be applied.
As shown in fig. 4, the system architecture 400 may include terminal devices 401, 402, 403, a network 404, and a server 405. The network 404 serves as a medium for providing communication links between the terminal devices 401, 402, 403 and the server 405. Network 404 may include various types of connections, such as wire, wireless communication links, or fiber optic cables, to name a few.
A user may use terminal devices 401, 402, 403 to interact with a server 405 over a network 404 to receive or send messages or the like. The terminal devices 401, 402, 403 may have installed thereon various communication client applications, such as shopping-like applications, web browser applications, search-like applications, instant messaging tools, mailbox clients, social platform software, etc. (by way of example only).
The terminal devices 401, 402, 403 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, and the like.
The server 405 may be a server providing various services, such as a background management server (for example only) providing support for shopping websites browsed by users using the terminal devices 401, 402, 403. The background management server can analyze and process the received data such as the product information inquiry request and feed back the processing result to the terminal equipment.
It should be noted that the method for item picking time duration prediction provided by the embodiments of the present invention is generally performed by the server 405, and accordingly, the means for item picking time duration prediction is generally disposed in the server 405.
It should be understood that the number of terminal devices, networks, and servers in fig. 4 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
Referring now to FIG. 5, shown is a block diagram of a computer system 500 suitable for use with a terminal device implementing an embodiment of the present invention. The terminal device shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.
As shown in fig. 5, the computer system 500 includes a Central Processing Unit (CPU)501 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)502 or a program loaded from a storage section 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data necessary for the operation of the system 500 are also stored. The CPU 501, ROM 502, and RAM 503 are connected to each other via a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.
The following components are connected to the I/O interface 505: an input portion 506 including a keyboard, a mouse, and the like; an output portion 507 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage portion 508 including a hard disk and the like; and a communication section 509 including a network interface card such as a LAN card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. The driver 510 is also connected to the I/O interface 505 as necessary. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as necessary, so that a computer program read out therefrom is mounted into the storage section 508 as necessary.
In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 509, and/or installed from the removable medium 511. The computer program performs the above-described functions defined in the system of the present invention when executed by the Central Processing Unit (CPU) 501.
It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules described in the embodiments of the present invention may be implemented by software or hardware. The described modules may also be provided in a processor, which may be described as: a processor includes a sample data acquisition module, a kernel matrix determination module, and a model determination module. The names of these modules do not in some cases constitute a limitation on the module itself, for example, the sample data acquisition module may also be described as a "module that acquires the characteristic data and the historical picking duration of the sample object".
As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be separate and not incorporated into the device. The computer readable medium carries one or more programs which, when executed by a device, cause the device to comprise: acquiring characteristic data and historical picking duration of a sample object; determining a kernel matrix based on a preset kernel function and the characteristic data of the sample object; and determining a picking duration prediction model according to the historical picking durations of the sample objects and the kernel matrix.
For the embodiment of the invention, the kernel function can expand the information with limited dimension into the information with infinite dimension, so the technical scheme solves the problem of insufficient characteristic information by using the kernel function. And based on the kernel function, the functional relation between the picking time and the characteristic data can be conveniently determined, and the problem that the existing functional model needs to assume a model form in advance is solved. The embodiment of the invention not only improves the accuracy of prediction and reduces the error rate, but also reduces the workload in logistics simulation.
The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (13)

1. A method for item picking duration prediction, comprising:
acquiring characteristic data and historical picking duration of a sample object;
determining a kernel matrix based on a preset kernel function and the characteristic data of the sample object;
and determining a picking duration prediction model according to the historical picking durations of the sample objects and the kernel matrix.
2. The method of claim 1, wherein the step of determining a kernel matrix based on a preset kernel function and the feature data of the sample object comprises:
mapping the characteristic data of the sample object to an infinite dimension based on a preset kernel function; the preset kernel function is a Gaussian kernel function;
and determining a kernel matrix according to the mapping result.
3. The method of claim 2, wherein the step of mapping the feature data of the sample object to infinite dimensions based on a preset kernel function comprises:
determining regulation and control parameters of a preset kernel function through ten-fold cross validation;
and mapping the characteristic data of the sample object to infinite dimensionality according to the regulation and control parameter and a preset kernel function.
4. The method of claim 1, wherein the step of determining a picking duration prediction model based on the historical picking durations of the sample objects and the kernel matrix comprises:
the model parameters α are determined by the following formula:
Figure FDA0001803457200000011
wherein, K is a nuclear matrix,
Figure FDA0001803457200000012
selecting the historical picking time length of the sample object;
based on the model parameters α, determining a picking time duration prediction model as:
y＝[1+k(x1,x),...,1+k(xn,x)]α
where y is the picking duration of the object, x1,...,xNCharacteristic data of the n sample objects respectively; k (x)iX) is a kernel function, i ═ 1, 2.
5. The method of claim 4, after building a picking duration prediction model based on historical picking durations for the sample objects and the kernel matrix, further comprising:
reducing the number of the feature data according to a preset step length, and taking the reduced feature data as a feature data set to be calculated;
determining the mean square error of each feature data set to be calculated based on the determined sorting duration prediction model;
taking the characteristic data set to be calculated corresponding to the minimum mean square error as an adjustment characteristic data set;
adjusting the picking length prediction model based on the adjusted feature data set.
6. The method according to claim 1, characterized in that the characteristic data of the sample object comprise at least: number of single picks, weight, volume, number of shelves on which the sample object is located.
7. An apparatus for item picking duration prediction, comprising:
the sample data acquisition module is used for acquiring the characteristic data and the historical picking duration of the sample object;
the kernel matrix determining module is used for determining a kernel matrix based on a preset kernel function and the characteristic data of the sample object;
and the model determining module is used for determining a picking duration prediction model according to the historical picking duration of the sample object and the nuclear matrix.
8. The apparatus of claim 7, wherein the kernel matrix determination module is further configured to map feature data of the sample object to infinite dimensions based on a preset kernel function; the preset kernel function is a Gaussian kernel function; and determining a kernel matrix according to the mapping result.
9. The apparatus of claim 8, wherein the kernel matrix determining module is further configured to determine a regulation parameter of a preset kernel function through ten-fold cross validation; and mapping the characteristic data of the sample object to infinite dimensionality according to the regulation and control parameter and a preset kernel function.
10. The apparatus of claim 7, wherein the model determination module is further configured to determine model parameters α by:
Figure FDA0001803457200000031
wherein, K is a nuclear matrix,
Figure FDA0001803457200000032
selecting the historical picking time length of the sample object;
based on the model parameters α, determining a picking time duration prediction model as:
y＝[1+k(x1,x),...,1+k(xn,x)]α
where y is the picking duration of the object, x1,...,xNIs respectively n samplesCharacteristic data of the subject; k (x)iX) is a kernel function, i ═ 1, 2.
11. The device according to claim 10, further comprising an adjusting module, configured to reduce the number of the feature data according to a preset step length, and use the reduced feature data as a feature data set to be calculated; determining the mean square error of each feature data set to be calculated based on the determined sorting duration prediction model; taking the characteristic data set to be calculated corresponding to the minimum mean square error as an adjustment characteristic data set; adjusting the picking length prediction model based on the adjusted feature data set.
12. An electronic device, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-6.
13. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-6.
CN201811087230.0A 2018-09-18 2018-09-18 Method and device for predicting item picking time Active CN110909908B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811087230.0A CN110909908B (en) 2018-09-18 2018-09-18 Method and device for predicting item picking time

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811087230.0A CN110909908B (en) 2018-09-18 2018-09-18 Method and device for predicting item picking time

Publications (2)

Publication Number Publication Date
CN110909908A true CN110909908A (en) 2020-03-24
CN110909908B CN110909908B (en) 2024-06-21

Family

ID=69813626

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811087230.0A Active CN110909908B (en) 2018-09-18 2018-09-18 Method and device for predicting item picking time

Country Status (1)

Country Link
CN (1) CN110909908B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113283826A (en) * 2021-03-23 2021-08-20 北京京东振世信息技术有限公司 Method and system for delivering articles out of warehouse
CN113537671A (en) * 2020-04-17 2021-10-22 北京京东振世信息技术有限公司 Sorting timeliness prediction method and device, storage medium and electronic equipment
CN115009744A (en) * 2022-06-23 2022-09-06 上海壹佰米网络科技有限公司 Method and device for estimating delivery time of front-end warehouse goods

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105469201A (en) * 2015-07-20 2016-04-06 浙江工业大学 Method for logistics dispensing center work task processing and scheduling
CN106021924A (en) * 2016-05-19 2016-10-12 华南理工大学 Sewage online soft-measurement method based on multi-attribute Gaussian kernel function fast relevance vector machine
CN106203737A (en) * 2016-07-29 2016-12-07 华北电力大学（保定） The multistep forecasting method of a kind of wind farm wind velocity and system
CN106682782A (en) * 2016-12-30 2017-05-17 国网新疆电力公司电力科学研究院 Short-term photovoltaic power prediction method based on EWT-KMPMR (empirical wavelet transform and kernel minimax probability machine classification)
CN106980955A (en) * 2017-03-29 2017-07-25 北京京东尚科信息技术有限公司 Method and apparatus for determining shelf storage space for shelf
CN107273921A (en) * 2017-05-31 2017-10-20 北京京东尚科信息技术有限公司 A kind of freight charges error prediction method and device based on SVM
CN107544944A (en) * 2017-09-04 2018-01-05 江西理工大学 A kind of SVMs Selection of kernel function method and its application based on graph theory
DE202018102632U1 (en) * 2018-05-09 2018-05-22 Robert Bosch Gmbh Device for creating a model function for a physical system
US20180152946A1 (en) * 2016-11-28 2018-05-31 Institute For Information Industry Channel selection device and channel selection method
CN108256586A (en) * 2018-01-31 2018-07-06 江西理工大学 A kind of support vector machines Selection of kernel function method and its application based on fractal theory

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105469201A (en) * 2015-07-20 2016-04-06 浙江工业大学 Method for logistics dispensing center work task processing and scheduling
CN106021924A (en) * 2016-05-19 2016-10-12 华南理工大学 Sewage online soft-measurement method based on multi-attribute Gaussian kernel function fast relevance vector machine
CN106203737A (en) * 2016-07-29 2016-12-07 华北电力大学（保定） The multistep forecasting method of a kind of wind farm wind velocity and system
US20180152946A1 (en) * 2016-11-28 2018-05-31 Institute For Information Industry Channel selection device and channel selection method
CN106682782A (en) * 2016-12-30 2017-05-17 国网新疆电力公司电力科学研究院 Short-term photovoltaic power prediction method based on EWT-KMPMR (empirical wavelet transform and kernel minimax probability machine classification)
CN106980955A (en) * 2017-03-29 2017-07-25 北京京东尚科信息技术有限公司 Method and apparatus for determining shelf storage space for shelf
CN107273921A (en) * 2017-05-31 2017-10-20 北京京东尚科信息技术有限公司 A kind of freight charges error prediction method and device based on SVM
CN107544944A (en) * 2017-09-04 2018-01-05 江西理工大学 A kind of SVMs Selection of kernel function method and its application based on graph theory
CN108256586A (en) * 2018-01-31 2018-07-06 江西理工大学 A kind of support vector machines Selection of kernel function method and its application based on fractal theory
DE202018102632U1 (en) * 2018-05-09 2018-05-22 Robert Bosch Gmbh Device for creating a model function for a physical system

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
SWARNENDU GHOSH 等: "The journey of graph kernels through two decades", 《COMPUTER SCIENCE REVIEW》, pages 88 - 111 *
吴倩红;高军;侯广松;韩蓓;汪可友;李国杰;: "实现影响因素多源异构融合的短期负荷预测支持向量机算法", 电力系统自动化, no. 15, pages 73 - 78 *
李林超;张健;杨帆;冉斌;: "基于核函数切换和支持向量回归的交通量短时预测模型", 东南大学学报(自然科学版), no. 05, pages 191 - 195 *

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113537671A (en) * 2020-04-17 2021-10-22 北京京东振世信息技术有限公司 Sorting timeliness prediction method and device, storage medium and electronic equipment
CN113283826A (en) * 2021-03-23 2021-08-20 北京京东振世信息技术有限公司 Method and system for delivering articles out of warehouse
CN113283826B (en) * 2021-03-23 2023-11-07 北京京东振世信息技术有限公司 Method and system for delivering articles out of warehouse
CN115009744A (en) * 2022-06-23 2022-09-06 上海壹佰米网络科技有限公司 Method and device for estimating delivery time of front-end warehouse goods
CN115009744B (en) * 2022-06-23 2024-06-04 上海壹佰米网络科技有限公司 Method and device for predicting delivery duration of front cargo

Also Published As

Publication number Publication date
CN110909908B (en) 2024-06-21

Similar Documents

Publication Publication Date Title
KR102170105B1 (en) Method and apparatus for generating neural network structure, electronic device, storage medium
CN110632608B (en) Target detection method and device based on laser point cloud
US9959670B2 (en) Method for rendering terrain
CN115131566A (en) Automatic image segmentation method based on super-pixels and improved fuzzy C-means clustering
CN110909908A (en) Method and device for predicting item picking duration
CN109960826B (en) Supply chain simulation method and device
US20220076169A1 (en) Federated machine learning using locality sensitive hashing
US11699094B2 (en) Automatic feature selection and model generation for linear models
CN113870334B (en) Depth detection method, device, equipment and storage medium
JP7270691B2 (en) Methods, apparatus, electronics, storage media and computer programs for labeling data
CN113743971A (en) Data processing method and device
CN110633717A (en) Training method and device for target detection model
CN110703758A (en) Path planning method and device
CN115495705A (en) Evaluation function determination method, evaluation function determination device, electronic device, and storage medium
CN115170815A (en) Method, device and medium for processing visual task and training model
CN109255563B (en) Method and device for determining storage area of article
US20220138557A1 (en) Deep Hybrid Graph-Based Forecasting Systems
CN113361739B (en) Method and device for generating picking path
CN113450042B (en) Method and device for determining goods supplementing quantity
CN115827914A (en) Method and device for generating relational topological graph, storage medium and electronic equipment
CN113822301B (en) Sorting center sorting method and device, storage medium and electronic equipment
US20220138786A1 (en) Artificial intelligence (ai) product including improved automated demand learning module
CN113361719A (en) Incremental learning method based on image processing model and image processing method
CN113538080A (en) Task list splitting method and device
CN112183799A (en) Task allocation method and device for synthesizing task list

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
TA01 Transfer of patent application right

Effective date of registration: 20200601

Address after: Room A1905, 19th floor, No. 2 Building, 18 Kechuang 11th Street, Beijing Daxing District, Beijing

Applicant after: Beijing Jingdong Qianshi Technology Co.,Ltd.

Address before: 300 457 days Tianjin Binhai New Area, Tianjin Economic and Technological Development Zone, No. 10 Fourth Avenue, Phase 2, First Floor

Applicant before: Tianjin Jingdong Shentuo Robot Technology Co.,Ltd.

TA01 Transfer of patent application right
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant