CN111126860A - Task allocation method, task allocation device and electronic equipment - Google Patents
Task allocation method, task allocation device and electronic equipment Download PDFInfo
- Publication number
- CN111126860A CN111126860A CN201911368857.8A CN201911368857A CN111126860A CN 111126860 A CN111126860 A CN 111126860A CN 201911368857 A CN201911368857 A CN 201911368857A CN 111126860 A CN111126860 A CN 111126860A
- Authority
- CN
- China
- Prior art keywords
- task
- operator
- information
- expression
- allocation
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 48
- 238000009826 distribution Methods 0.000 claims abstract description 52
- 238000007781 pre-processing Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 16
- 238000012805 post-processing Methods 0.000 claims description 13
- 238000004590 computer program Methods 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 8
- 230000010365 information processing Effects 0.000 claims description 6
- 239000013598 vector Substances 0.000 description 17
- 239000011159 matrix material Substances 0.000 description 14
- 230000008569 process Effects 0.000 description 10
- 238000004891 communication Methods 0.000 description 9
- 238000010586 diagram Methods 0.000 description 4
- 239000000284 extract Substances 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000013507 mapping Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 239000000654 additive Substances 0.000 description 1
- 230000000996 additive effect Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000005303 weighing Methods 0.000 description 1
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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
-
- 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
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0631—Resource planning, allocation, distributing or scheduling for enterprises or organisations
- G06Q10/06311—Scheduling, planning or task assignment for a person or group
- G06Q10/063112—Skill-based matching of a person or a group to a task
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Entrepreneurship & Innovation (AREA)
- Educational Administration (AREA)
- Operations Research (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Game Theory and Decision Science (AREA)
- Quality & Reliability (AREA)
- Tourism & Hospitality (AREA)
- Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Machine Translation (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a task allocation method, a task allocation device and electronic equipment, wherein the task allocation method comprises the following steps: determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics; and inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model. The task allocation method provided by the embodiment of the invention performs task allocation based on the ideas of feature modeling and feature matching, has high feature integrity, and can improve the matching degree of task allocation.
Description
Technical Field
The present invention relates to the field of task allocation, and in particular, to a task allocation method, a task allocation apparatus, and an electronic device.
Background
The task allocation is used for establishing a mapping relation between the task to be completed and the operator to receive the task, and the accuracy of pairing between the task and the operator directly influences the execution efficiency and the execution effect of all tasks.
The conventional task allocation method comprises the steps of firstly modeling a mathematical problem such as a linear programming problem, a dynamic programming problem and the like according to a problem, then solving the programming problem by adopting a corresponding traditional thought according to the problem type, and finally outputting an optimal result obtained by the solution.
In other words, in the modeling and counting process, approximation and simplification processing can be continuously performed, problems are artificially and continuously simplified in the whole distribution process, the problems and characteristics obtained by modeling have larger difference with actual scenes, and finally the inconsistency of problem modeling and actual is caused, and the reasonability of distribution results, the maximization of resource utilization and the like can be greatly reduced.
Disclosure of Invention
Embodiments of the present invention provide a task allocation method, a task allocation apparatus, and an electronic device that overcome the above problems or at least partially solve the above problems.
In a first aspect, an embodiment of the present invention provides a task allocation method, including: determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics; inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model; the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: the content expression characteristic of the implicit characteristic of the sample task and the predetermined sample task type quantity characteristic, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
In some embodiments, the inputting the task information to be distributed and the operator information of the task to be received into a task distribution model to obtain a task distribution result output by the task distribution model includes: inputting the content expression characteristics of the task recessive characteristics, the task dominant characteristics, the content expression characteristics of the operator recessive characteristics and the operator dominant characteristics to a pre-processing layer of the task allocation model to obtain the task expression characteristics and the operator expression characteristics; inputting the task expression characteristics and the operator expression characteristics to a task allocation layer of the task allocation model to obtain task allocation expression characteristics; and inputting the task allocation expression characteristics to a post-processing layer of the task allocation model to obtain the task allocation result.
In some embodiments, the inputting the content expression features of the task implicit features, the task explicit features, the content expression features of the operator implicit features, and the operator explicit features into a pre-processing layer of the task assignment model to obtain the task expression features and the operator expression features specifically includes: inputting the content expression characteristics of the task implicit characteristics to an input layer of the pre-processing layer to obtain the task implicit characteristics; inputting the content expression characteristics of the operator recessive characteristics into an input layer of the preprocessing layer to obtain the operator recessive characteristics; inputting the task-dominant features and the operator-dominant features into an auxiliary input layer of the pre-processing layer for feature coding; determining the task expression characteristics based on the task implicit characteristics; determining the operator expression signature based on the operator recessive signature and the predetermined operator dominant signature.
In some embodiments, the determining the task expression characteristics based on the task implicit characteristics includes: inputting the task implicit characteristics into a coding layer of the preprocessing layer to obtain the task expression characteristics; the determining the operator expression signature based on the operator recessive signature and the operator dominant signature includes: and inputting the operator recessive characteristic and the operator dominant characteristic into an encoding layer of the preprocessing layer to obtain the operator expression characteristic.
In some embodiments, the inputting the task expression features and the operator expression features into a task allocation layer of the task allocation model to obtain the task allocation expression features specifically includes: a primary distribution step: inputting the task expression characteristics and the operator expression characteristics into a task matching layer of the task distribution layer to obtain task matching representation, and updating the operator expression characteristics; and task adjustment: and inputting the task matching representation into an allocation adjustment layer of the task allocation layer to obtain the task allocation expression characteristics.
In some embodiments, if it is obtained that any operator is assigned multiple tasks based on the task matching representation, an adjustment is made based on the operator-explicit characteristics to enable any operator to assign a single category of tasks.
In some embodiments, if it is known based on the task allocation expression features that any task does not determine a corresponding operator, the preliminary allocation step is repeated with the task allocation expression features and the updated operator expression features as inputs to the task matching layer; and if any operator determines the corresponding operator based on the task allocation expression characteristics, taking the task allocation expression characteristics as the output of the task allocation layer.
In some embodiments, the inputting the task allocation expression characteristics into a post-processing layer of the task allocation model to obtain the task allocation result includes: inputting the task allocation expression characteristics to a decoding layer of the post-processing layer to obtain the allocation probability of the operator corresponding to the task; and inputting the distribution probability and the task type quantity characteristics to an output layer of the post-processing layer to obtain the task distribution result, wherein the task explicit characteristics comprise the task type quantity characteristics.
In a second aspect, an embodiment of the present invention provides a task allocation apparatus, including: the system comprises an information determining unit, a task information processing unit and a task information processing unit, wherein the information processing unit is used for determining task information to be distributed and operator information of a task to be received, the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics; the task allocation unit is used for inputting the task information to be allocated and the operator information of the task to be received into a task allocation model and obtaining a task allocation result output by the task allocation model; the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of the sample task implicit characteristics and predetermined sample task explicit characteristics, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
In a third aspect, an embodiment of the present invention provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the steps of the method provided in the first aspect when executing the program.
In a fourth aspect, an embodiment of the present invention provides a non-transitory computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the method as provided in the first aspect.
The task allocation method, the task allocation device and the electronic equipment of the embodiment of the invention perform task allocation based on the ideas of feature modeling and feature matching, have high feature integrity and can improve the matching precision of task allocation.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flow chart of a task allocation method according to an embodiment of the present invention;
FIG. 2 is a schematic structural diagram of a task allocation model according to an embodiment of the present invention;
FIG. 3 is a schematic structural diagram of a task allocation apparatus according to an embodiment of the present invention;
fig. 4 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A task allocation method of an embodiment of the present invention is described below with reference to fig. 1-2.
As shown in fig. 1, the task allocation method according to the embodiment of the present invention includes the following steps:
step S100, determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of the task implicit characteristics and predetermined task explicit characteristics, and the operator information comprises content expression characteristics of the operator implicit characteristics and predetermined operator explicit characteristics.
The information determined as described above is divided into a task side and an operator side.
The task information of the task side comprises the following steps: the content of the task implicit features expresses the features and the predetermined task explicit features. The content expression characteristics of the task implicit characteristics are implicit task implicit characteristics, the content expression characteristics of the task implicit characteristics can be task introduction content or other information related to the task, the task implicit characteristics are implicit in the information, the task implicit characteristics are not easy to discover or arrange manually, for example, the task implicit characteristics can include at least one of the following three types: (1) manually undiscovered features associated with task execution requirements; (2) manually discovered information that is related to task performance requirements, but is difficult to quantify or count; (3) information that can be counted but the statistical results are not sufficiently accurate. In practical implementation, the content expression feature of the task implicit feature can be in a coded form, different task types are respectively represented by different vectors, and the vector dimension is 128.
The predetermined task-dominant features are manually organized information, for example, the task-dominant features may include task-type quantity features, and the task-type quantity features may express the features in the form of vectors, and have a length of 32, so that the task-type quantity features are more complete, in other words, the task-type quantity features may characterize the number of tasks of each type. This feature is a mission-dominant feature.
In some embodiments, the task-explicit feature may include, in addition to the task-type-quantity feature: the task experience characteristics are determined according to the manual priori knowledge, the characteristics can be task dominant characteristics, the task dominant characteristics comprise known requirements of the task on the capability of an operator, background information of the task and the like, if 10 aspects of the task requirement information are known, a 10-dimensional vector is added for coding, if one dimension is the requirement that the operator has a better English level, the English level can be graded, such as 1 grade, 2 grade, 3 grade and the like, the requirements of the task on the English level of the operator are represented by 1, 2 and 3 respectively, and task description information is also the same.
The operator information on the operator side includes: the content of the operator recessive feature expresses the feature and the predetermined operator dominant feature.
The content expression characteristic of the operator recessive characteristic implicitly includes the operator recessive characteristic, and the content expression characteristic of the operator recessive characteristic may be operator history information related to an operator, in which the operator recessive characteristic is implicitly included, and the operator recessive characteristic is not easily discovered or collated by a human, for example, the operator recessive characteristic may include at least one of the following three types: (1) manually undiscovered features associated with performing the task; (2) manually discovered information that is relevant to performing a task, but difficult to quantify or count; (3) information that can be counted but the statistical results are not sufficiently accurate. In practical implementation, the content-expressing feature of the operator-implicit feature may be in an encoded form, and different task types are respectively represented by different vectors, and the vector dimension is 128.
The predetermined operator dominant features are information which is manually arranged, the operator dominant features can be coded for the known ability of the operator (such as the ability of task completion, self academic history, skill and the like), vectors of corresponding dimensions are added for coding according to the ability types, and if the academic requirements are divided into doctor, master, subject, special subject, high school and low school, the coding is carried out on the dimension of the academic history according to the academic levels of different people. In addition, additional operator information such as acceptable task number codes may need to be added.
And S200, inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model.
The task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises the following steps: the content expression characteristic of the sample task implicit characteristic and the predetermined sample task explicit characteristic, and the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
In other words, the training samples of the task assignment model are: sample task information and sample operator information; the training labels of the task allocation model are as follows: and a task allocation result, which can be obtained by manually marking according to the sample task information and the sample operator information.
It can be understood that the parameters considered by the task assignment model include task-explicit and task-implicit features and operator-explicit and operator-implicit features, which ensure the integrity of the model input information. Compared with pure artificial feature modeling in the related technology, the problem that the model is inconsistent with the actual situation can be solved.
The task allocation method provided by the embodiment of the invention performs task allocation based on the ideas of feature modeling and feature matching, has high feature integrity, and can improve the matching precision of task allocation.
In some embodiments, step S200, inputting task information to be distributed and operator information of a task to be received into the task distribution model, and obtaining a task distribution result output by the task distribution model, includes: step S210, step S220, and step S230.
In step S210, the content expression characteristics of the task implicit characteristics, the task explicit characteristics, the content expression characteristics of the operator implicit characteristics, and the predetermined operator explicit characteristics are input to the pre-processing layer of the task allocation model, so as to obtain the task expression characteristics and the operator expression characteristics.
As shown in fig. 2, a pre-processing layer of the task allocation model extracts the task implicit features from the content expression features of the task implicit features to generate task expression features, and tasks with the same and similar attributes in the task expression features can obtain similar vectors in the space of the tasks, which is beneficial to subsequent task allocation.
Certainly, the pre-processing layer of the task allocation model can also extract the task implicit features in the content expression features of the task implicit features, and then generate the task expression features by combining the task explicit features.
The pre-processing layer of the task allocation model extracts the operator recessive features in the content expression features of the operator recessive features, the operator expression features are generated by combining the operator dominant features, operators with the same and similar attributes in the operator expression features can obtain similar vectors in the space of the operators, and subsequent task allocation is facilitated.
Further, step S210, inputting the content expression features of the task implicit features, the task explicit features, the content expression features of the operator implicit features, and the predetermined operator explicit features into a pre-processing layer of the task allocation model to obtain the task expression features and the operator expression features, which specifically includes: step S211, step S212, and step S213.
And S211, inputting the content expression characteristics of the task implicit characteristics to an input layer of a preprocessing layer to obtain the task implicit characteristics.
And S212, inputting the content expression characteristics of the operator recessive characteristics into an input layer of the preprocessing layer to obtain the operator recessive characteristics.
The input layer learns autonomously in a large amount of training data through a task allocation model to obtain various characteristics of tasks and operators.
The input layer comprises a task output coding module and an operator input coding module, and the task output module is used for determining the task implicit characteristic according to the content expression characteristic of the task implicit characteristic; the operator input module is used for determining the recessive characteristic of the operator according to the content expression characteristic of the recessive characteristic of the operator.
In other words, the input layer encodes information by two matrix input different encoding modules (a task output encoding module and an operator input encoding module).
In the training process, the manually distributed historical records are used as supervision signals, in the training and optimizing process, the features of the tasks and the operators are continuously captured in the training process through gradient feedback and updating, the feature representations of the tasks and the operators are continuously updated, and the task implicit features and the operator implicit features are finally obtained. In addition, the intensive learning of the model can be realized by using the evaluation value of the evaluation function as a supervision signal of the task allocation model.
And step S213, inputting the predetermined task dominant characteristic and the predetermined operator dominant characteristic into an auxiliary input layer of the pre-processing layer for characteristic coding.
The auxiliary input layer is used for inputting the priori knowledge of manual arrangement and some characteristics which are helpful for understanding and coding the task and the human ability, so that the model is easier to learn and train. The auxiliary input layer may also input task dominant features.
In other words, as shown in FIG. 2, the task assignment model includes: an input layer and an auxiliary input layer. The input layer is used for determining the task implicit characteristic according to the content expression characteristic of the task implicit characteristic, and the input layer is also used for determining the operator implicit characteristic according to the content expression characteristic of the operator implicit characteristic; the auxiliary input layer is used for inputting a predetermined task type quantity characteristic and a predetermined operator characteristic.
And step S214, determining task expression characteristics based on the task implicit characteristics.
And S215, determining an operator expression characteristic based on the operator recessive characteristic and a predetermined operator dominant characteristic.
Further, in step S214, determining a task expression feature based on the task implicit feature, including: inputting the task implicit characteristics into a coding layer of a preprocessing layer to obtain task expression characteristics;
step S215, determining an operator expression characteristic based on the operator implicit characteristic and the predetermined operator explicit characteristic, wherein the method comprises the following steps: and inputting the recessive characteristic of the operator and the predetermined dominant characteristic of the operator into an encoding layer of a preprocessing layer to obtain the expression characteristic of the operator.
In other words, as shown in fig. 2, the task assignment model further includes: the system comprises a coding layer and an operator expression layer, wherein the coding layer is used for determining task expression characteristics according to task recessive characteristics and task dominant characteristics, and the coding layer is also used for determining operator expression characteristics according to operator recessive characteristics and operator dominant characteristics.
It is to be understood that the encoding layer may include a task encoding module and an operator encoding module.
The task coding module is used for determining relevant information between tasks, and the operator coding module is used for determining relevant information between operators. The task coding module and the operator coding module can adopt a Transformer translation model for coding, the Transformer consists of 6 coding layers, each layer comprises a multi-head attention module, a connection module, a residual connection module and the like, and the characteristic representation of each task and each operator is obtained after coding. The depth of the neural network is deepened through a plurality of layers of coding modules, so that the extraction of the features of higher layers is facilitated, and the effect of the model is improved.
And S220, inputting the task expression characteristics and the operator expression characteristics into a task allocation layer of the task allocation model to obtain the task allocation expression characteristics.
The task assignment expression feature is used to represent the probability that each task is assigned to the corresponding operator.
In other words, as shown in FIG. 2, the task assignment model includes a task assignment layer for determining task assignment expression characteristics based on the task expression characteristics and the operator expression characteristics.
Further, step S220, inputting the task expression features and the operator expression features into a task allocation layer of the task allocation model to obtain the task allocation expression features, which specifically includes: step S221 and step S222.
Step S221, preliminary allocation step: and inputting the task expression characteristics and the operator expression characteristics into a task matching layer of the task distribution layer to obtain task matching representation, and updating the operator expression characteristics.
The task matching representation can preliminarily determine the matching relationship of the task and the operator.
The task distribution layer comprises a task matching layer, and the task matching layer is used for determining task matching representation and updated operator expression characteristics according to the task expression characteristics and the operator expression characteristics.
The task distribution layer fuses matching results into coding of tasks through characteristic matching between the tasks and operators (distribution objects), so that balance and exchange of distribution information of the tasks are carried out in the next layer (distribution adjustment layer). For example, the characteristics of task a and operator B are more matched, in the module, through mutual matching between task and allocation object, B is arranged to complete task a, that is, information related to B is superimposed on the characteristics of a, so that a has preliminary allocation result information, and the characteristics of task a at this time include both the inherent characteristic information of a and the allocation result information of task a.
The method is implemented in practical implementation as follows:
(1) task feature matrix x obtained for previous layer(n-1)' and operator characteristics matrix y(n-1)Linear transformed and operator featuresMatching and overlapping, in other words, performing additive attention operation, and calculating a correlation coefficient W between the task x and the operator y to obtain a matching matrix:
wherein x is a task coding matrix, y is a personnel information coding matrix, and W1, W2 and b are optimization parameters of feature fusion. V1 TAs a coordinate mapping matrix, A1The normalized correlation matrix represents the matching degree of the task x and the operator y, namely the matching degree of each task and each operator.
(2) After the matching matrix is normalized through a softmax function according to rows, the matching weight W' between the task and the operator is obtained, when the matching weight is higher, the task is related to the characteristics of the operator most, namely the task is suitable for being completed by the operator, and the operator is a candidate operator:
W'＝softmax(A1)
(3) and (3) carrying out weighted summation on the operator characteristics according to the weight matrix obtained in the step (2) to obtain a candidate operator information representation vector related to the task. And the obtained vector is superposed on the characteristics of the original task after linear transformation, so that the task-related operator is distributed to the current task.
Obtaining the final task allocation expression x by balancing the task of the multi-layer mutual attention layer with the characteristics of operators and the information exchange and allocation result among the tasks of the multi-layer self-attention layer(n)：
x(n)＝W'y(n-1)+x(n-1)′
Task assignment representation x(n)It may be preliminarily determined to which operators each task is assigned.
(4) And updating the operator expression characteristics to obtain the distributed state updating characteristics.
y (n)＝softmax(A2)x(n-1)'+y (n-1)
Wherein, V2 TTransforming the matrix for features, A2For the correlation matrix mapped to the task space, A2For updating the information of the operators, i.e. which operators have not yet obtained the allocation. y is(n)The coded information mainly comprises self characteristics, allocated personnel information and residual allocable capacity information for a new characteristic matrix of an operator after one round of allocation. The capacity is adjusted according to the number of tasks that have been allocated.
Step S222, task adjustment: and inputting the task matching expression into an allocation adjustment layer of the task allocation layer to obtain the task allocation expression characteristics.
The task allocation layer comprises an allocation adjustment layer, and the allocation adjustment layer is used for determining the task allocation expression characteristics according to the task matching expression.
Further, if it is acquired that any operator is assigned with a plurality of types of tasks based on the task matching indication, adjustment is performed based on the operator-dominant feature so that any operator is assigned with a single type of task. If it is determined that any operator is not assigned multiple tasks based on the task matching, in other words, any operator is assigned a single task, no adjustment is performed.
The distribution adjusting layer highlights the relation between tasks by fusing information between different tasks, vectors after task coding with similar types are more similar in the space, and meanwhile, the vectors are distinguished according to task priority difference and quantity difference. Through the mutual matching layer, originally tasks A, B of the same type are more similar, but because the tasks are successfully matched with operators with similar capabilities in the task matching layer, in the module, the tasks which obtain the same pairing operator mark are mutually coordinated according to the operator information obtained by the matching layer, so that information exchange among different tasks and balance of a distribution result are realized, a comparison and balance process of a human to a complex problem is simulated, and the distribution result is more reasonable. If both task a and task B are assigned to operator a, but the operator has limited time and energy and cannot complete too many tasks, in this layer, after mutual weighing of the assignment results, task a is assigned to operator a, and task B is to retrieve a new operator in the next round of matching.
The method is implemented in practical implementation as follows:
(1) task features x to task matching layer(n)Making a non-linear transformation of sigma (W) in each dimension3x(n)+b1) Wherein W is3And b1Transformation parameters for model training and optimization;
(2) introducing an external variable U, and performing inner product with each dimension in (1) to obtain the weight of the control input, namely:
B＝UTσ(W3x(n)+b1)
(3) performing normalization operation on the vector after inner product in the step (2), and performing weight distribution, namely:
weight＝softmax(B)
(4) and (4) carrying out linear combination on the input matrixes by using the weights in the step (3) as a new round of expression characteristic of the task:
x(n)′＝weight·x(n)
in some embodiments, if it is known based on the task allocation expression features that any task does not determine a corresponding operator, the preliminary allocation step is repeated with the task allocation expression features and the updated operator expression features as inputs to the task matching layer. And if any operator determines the corresponding operator based on the task allocation expression characteristics, taking the task allocation expression characteristics as the output of the task allocation layer.
It will be appreciated that the task allocation layer does not complete the allocation until all tasks are matched to the corresponding operators.
In other words, the task allocation layer includes multiple layers of pre-allocation and multi-layer adjustment, divided into a task matching layer and an allocation adjustment layer. And the final personnel allocation is realized through the processes of multi-layer pre-allocation and readjustment.
And step S230, inputting the task allocation expression characteristics to a post-processing layer of the task allocation model to obtain a task allocation result.
In other words, the task allocation model includes a post-processing layer for determining the task allocation result according to the task allocation expression characteristics, as shown in fig. 2.
In some embodiments, step S230, inputting the task allocation expression feature into a post-processing layer of the task allocation model, and obtaining a task allocation result includes: step S231 and step S232.
And S231, inputting the task allocation expression characteristics to a decoding layer of the post-processing layer to obtain the allocation probability of the operator corresponding to the task.
In other words, as shown in FIG. 2, the task assignment model includes a decoding layer for determining an assignment probability that an operator corresponds to a task based on the task assignment expression characteristics.
In actual implementation, the decoding layer may use LSTM as a decoder for decoding to enhance mutual coordination in the task allocation process. The LSTM comprises an input gate, an output gate and a memory gate, in the decoding process, decoding is carried out in sequence according to the input sequence of tasks, and the subsequent decoding tasks can obtain the previously decoded information through the memory module of the LSTM, so as to dynamically adjust the decoding result. LSTM is horizontal decoding, that is, each task is decoded one by one, so if there is misallocation, after adjustment, the information will be transmitted backward through the feature stream (ht-1 and ct-1) of LSTM, in decoding, the input gate obtains the current allocation information, and the memory gate obtains the already decoded allocation result.
The current distribution information and the information of the memory gate are fused and output, so that the mutual coordination of resources is achieved, the output result is the adjusted result, and the current information is transmitted to the next LSTM through the state information c, namely the information input of the memory gate. The calculation process is as follows:
ht，ct＝LSTM([ht-1，)yt-1]，ct-1)
wherein the input of the LSTM comprises a hidden layer representation h of the moment immediately preceding itt-1State quantity ofct-1And the previous time output yt-1，[a,b]Showing the concatenation of a and b.
And mapping the decoded feature vector to a final output space after obtaining the decoded feature vector, and performing normalization processing on the decoded feature vector through a softmax function to obtain the distribution probability of each person corresponding to the task. The calculation is as follows:
yt＝softmax(W4ht+b2)
wherein, W4And b2Is a non-shared parameter.
And step S232, inputting the distribution probability and the predetermined task type quantity characteristics into an output layer of the post-processing layer to obtain a task distribution result, wherein the task explicit characteristics comprise the task type quantity characteristics.
In other words, as shown in fig. 2, the task allocation model includes an output layer, and the output layer is used for determining the task allocation result according to the allocation probability and the task type quantity characteristics.
The task assignment result is used to indicate to which operator each task is assigned, and several are assigned.
The decoding layer outputs a probability matrix for each task assigned to an operator. In actual implementation, the output layer is configured to multiply the allocation probability by the total number of tasks and perform rounding to obtain a final allocation result:
O＝[Cy+0.5]
wherein, O is the output result, y is the distribution probability, C is the task type quantity characteristic, i.e., the quantity of each task, [ +0.5] represents rounding.
The task allocation device provided by the embodiment of the present invention is described below, and the task allocation device described below and the task allocation method described above may be referred to correspondingly.
As shown in fig. 3, a task assigning apparatus according to an embodiment of the present invention includes: an information determination unit 510 and a task allocation unit 520.
The information determining unit 510 is configured to determine task information to be distributed and operator information of a task to be received, where the task information includes a content expression feature of a task implicit feature and a predetermined task explicit feature, and the operator information includes a content expression feature of an operator implicit feature and a predetermined operator explicit feature; the task allocation unit 520 is configured to input the task information to be allocated and the operator information of the task to be received into a task allocation model, and obtain a task allocation result output by the task allocation model; the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of the sample task implicit characteristics and predetermined sample task explicit characteristics, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
Fig. 4 illustrates a physical structure diagram of an electronic device, which may include, as shown in fig. 4: a processor (processor)810, a communication Interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication Interface 820 and the memory 830 communicate with each other via the communication bus 840. The processor 810 may call logic instructions in the memory 830 to perform a task assignment method comprising: determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics; inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model; the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of the sample task implicit characteristics and predetermined sample task explicit characteristics, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
It should be noted that, when being implemented specifically, the electronic device in this embodiment may be a server, a PC, or other devices, as long as the structure includes the processor 810, the communication interface 820, the memory 830, and the communication bus 840 shown in fig. 4, where the processor 810, the communication interface 820, and the memory 830 complete mutual communication through the communication bus 840, and the processor 810 may call the logic instructions in the memory 830 to execute the above method. The embodiment does not limit the specific implementation form of the electronic device.
In addition, the logic instructions in the memory 830 may be implemented in software functional units and stored in a computer readable storage medium when the logic instructions are sold or used as independent products. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
Further, an embodiment of the present invention discloses a computer program product, the computer program product includes a computer program stored on a non-transitory computer readable storage medium, the computer program includes program instructions, when the program instructions are executed by a computer, the computer can execute the task allocation method provided by the above method embodiments, the method includes: determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics; inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model; the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of the sample task implicit characteristics and predetermined sample task explicit characteristics, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
In another aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored, where the computer program is implemented to perform the task allocation method provided in the foregoing embodiments when executed by a processor, and the method includes: determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics; inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model; the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of the sample task implicit characteristics and predetermined sample task explicit characteristics, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A task allocation method, comprising:
determining task information to be distributed and operator information of a task to be received, wherein the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics;
inputting the task information to be distributed and the operator information of the task to be received into a task distribution model, and obtaining a task distribution result output by the task distribution model;
the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of the sample task implicit characteristics and predetermined sample task explicit characteristics, wherein the sample operator information comprises: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
2. The task allocation method according to claim 1, wherein the inputting the task information to be allocated and the operator information of the task to be received into a task allocation model to obtain a task allocation result output by the task allocation model comprises:
inputting the content expression characteristics of the task recessive characteristics, the task dominant characteristics, the content expression characteristics of the operator recessive characteristics and the operator dominant characteristics to a pre-processing layer of the task allocation model to obtain the task expression characteristics and the operator expression characteristics;
inputting the task expression characteristics and the operator expression characteristics to a task allocation layer of the task allocation model to obtain task allocation expression characteristics;
and inputting the task allocation expression characteristics to a post-processing layer of the task allocation model to obtain the task allocation result.
3. The task allocation method according to claim 2, wherein the inputting the content expression features of the task implicit features, the task explicit features, the content expression features of the operator implicit features, and the operator explicit features into a pre-processing layer of the task allocation model to obtain the task expression features and the operator expression features specifically comprises:
inputting the content expression characteristics of the task implicit characteristics to an input layer of the pre-processing layer to obtain the task implicit characteristics;
inputting the content expression characteristics of the operator recessive characteristics into an input layer of the preprocessing layer to obtain the operator recessive characteristics;
inputting the task-dominant features and the operator-dominant features into an auxiliary input layer of the pre-processing layer for feature coding;
determining the task expression characteristics based on the task implicit characteristics and the task explicit characteristics;
determining the operator expression signature based on the operator recessive signature and the operator dominant signature.
4. The task allocation method according to claim 3, wherein the determining the task expression characteristic based on the task implicit characteristic comprises: inputting the task implicit characteristics into a coding layer of the preprocessing layer to obtain the task expression characteristics;
the determining the operator expression signature based on the operator recessive signature and the operator dominant signature includes: and inputting the operator recessive characteristic and the operator dominant characteristic into an encoding layer of the preprocessing layer to obtain the operator expression characteristic.
5. The task allocation method according to any one of claims 2 to 4, wherein the inputting the task expression features and the operator expression features into a task allocation layer of the task allocation model to obtain task allocation expression features specifically comprises:
a primary distribution step: inputting the task expression characteristics and the operator expression characteristics into a task matching layer of the task distribution layer to obtain task matching representation, and updating the operator expression characteristics;
and task adjustment: and inputting the task matching representation into an allocation adjustment layer of the task allocation layer to obtain the task allocation expression characteristics.
6. The method according to claim 5, wherein if it is acquired that any operator is assigned multiple tasks based on the task matching representation, an adjustment is made based on the operator-explicit feature so that any operator is assigned a single type of task.
7. The task allocation method according to any one of claims 2 to 4, wherein the inputting the task allocation expression characteristics into a post-processing layer of the task allocation model to obtain the task allocation result comprises:
inputting the task allocation expression characteristics to a decoding layer of the post-processing layer to obtain the allocation probability of the operator corresponding to the task;
and inputting the distribution probability and the task type quantity characteristics to an output layer of the post-processing layer to obtain the task distribution result, wherein the task explicit characteristics comprise the task type quantity characteristics.
8. A task assigning apparatus, comprising:
the system comprises an information determining unit, a task information processing unit and a task information processing unit, wherein the information processing unit is used for determining task information to be distributed and operator information of a task to be received, the task information comprises content expression characteristics of task recessive characteristics and predetermined task dominant characteristics, and the operator information comprises content expression characteristics of operator recessive characteristics and predetermined operator dominant characteristics;
the task allocation unit is used for inputting the task information to be allocated and the operator information of the task to be received into a task allocation model and obtaining a task allocation result output by the task allocation model;
the task allocation model is obtained by taking sample task information and sample operator information as samples and taking task allocation results corresponding to the sample task information and the sample operator information as sample labels through training, wherein the sample task information comprises: content expression characteristics of sample task implicit characteristics and predetermined sample arbitrary characteristics, the sample operator information including: a content representative characteristic of the sample operator recessive characteristic and a predetermined sample operator dominant characteristic.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the task assigning method according to any of claims 1 to 7 are implemented when the processor executes the program.
10. A non-transitory computer readable storage medium, on which a computer program is stored, the computer program, when being executed by a processor, implementing the steps of the task allocation method according to any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911368857.8A CN111126860B (en) | 2019-12-26 | 2019-12-26 | Task allocation method, task allocation device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911368857.8A CN111126860B (en) | 2019-12-26 | 2019-12-26 | Task allocation method, task allocation device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111126860A true CN111126860A (en) | 2020-05-08 |
| CN111126860B CN111126860B (en) | 2023-12-26 |
Family
ID=70503374
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911368857.8A Active CN111126860B (en) | 2019-12-26 | 2019-12-26 | Task allocation method, task allocation device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111126860B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111738409A (en) * | 2020-05-14 | 2020-10-02 | 华为技术有限公司 | Resource scheduling method and related equipment thereof |
| CN112860416A (en) * | 2021-04-25 | 2021-05-28 | 城云科技（中国）有限公司 | Annotating task assignment strategy method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150254596A1 (en) * | 2014-03-07 | 2015-09-10 | Netflix, Inc. | Distributing tasks to workers in a crowd-sourcing workforce |
| CN109190929A (en) * | 2018-08-14 | 2019-01-11 | 中国平安人寿保险股份有限公司 | A kind of workload methodology, device, computer equipment and storage medium |
| CN109902901A (en) * | 2017-12-11 | 2019-06-18 | 广东智动力知识产权运营有限公司 | Method for allocating tasks, device, storage medium and computer equipment |
-
2019
- 2019-12-26 CN CN201911368857.8A patent/CN111126860B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150254596A1 (en) * | 2014-03-07 | 2015-09-10 | Netflix, Inc. | Distributing tasks to workers in a crowd-sourcing workforce |
| CN109902901A (en) * | 2017-12-11 | 2019-06-18 | 广东智动力知识产权运营有限公司 | Method for allocating tasks, device, storage medium and computer equipment |
| CN109190929A (en) * | 2018-08-14 | 2019-01-11 | 中国平安人寿保险股份有限公司 | A kind of workload methodology, device, computer equipment and storage medium |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111738409A (en) * | 2020-05-14 | 2020-10-02 | 华为技术有限公司 | Resource scheduling method and related equipment thereof |
| CN111738409B (en) * | 2020-05-14 | 2024-05-24 | 华为技术有限公司 | Resource scheduling method and related equipment thereof |
| CN112860416A (en) * | 2021-04-25 | 2021-05-28 | 城云科技（中国）有限公司 | Annotating task assignment strategy method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111126860B (en) | 2023-12-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112784092B (en) | Cross-modal image text retrieval method of hybrid fusion model | |
| CN110781406B (en) | Social network user multi-attribute inference method based on variational automatic encoder | |
| EP4350572A1 (en) | Method, apparatus and system for generating neural network model, devices, medium and program product | |
| WO2019242627A1 (en) | Data processing method and apparatus | |
| CN113628059A (en) | Associated user identification method and device based on multilayer graph attention network | |
| CN111784699B (en) | Method and device for carrying out target segmentation on three-dimensional point cloud data and terminal equipment | |
| CN111126860B (en) | Task allocation method, task allocation device and electronic equipment | |
| CN115130536A (en) | Training method of feature extraction model, data processing method, device and equipment | |
| CN116362325A (en) | Electric power image recognition model lightweight application method based on model compression | |
| CN114067308A (en) | Intelligent matching method and device, electronic equipment and storage medium | |
| CN109214515A (en) | A kind of deep neural network inference method and calculate equipment | |
| CN114677535A (en) | Training method of domain-adaptive image classification network, image classification method and device | |
| CN111898544A (en) | Character and image matching method, device and equipment and computer storage medium | |
| CN116489038A (en) | Network traffic prediction method, device, equipment and medium | |
| CN111079930A (en) | Method and device for determining quality parameters of data set and electronic equipment | |
| CN116958027A (en) | Three-dimensional industrial abnormality detection method and device, storage medium and electronic equipment | |
| CN118035800A (en) | Model training method, device, equipment and storage medium | |
| CN113642627B (en) | Deep learning-based image and decision multi-source heterogeneous information fusion identification method and device | |
| CN115577797A (en) | Local noise perception-based federated learning optimization method and system | |
| CN115982634A (en) | Application program classification method and device, electronic equipment and computer program product | |
| CN115063374A (en) | Model training method, face image quality scoring method, electronic device and storage medium | |
| CN110889635B (en) | Method for performing emergency drilling on food safety event processing | |
| CN114756425A (en) | Intelligent monitoring method and device, electronic equipment and computer readable storage medium | |
| CN114238106A (en) | Test time prediction method and device, electronic device and storage medium | |
| CN109308565B (en) | Crowd performance grade identification method and device, storage medium and computer equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |