WO2023035526A1 - Object sorting method, related device, and medium - Google Patents

Abstract

An object sorting method, a related device, and a medium, which are applied to the technical field of artificial intelligence. The method comprises: acquiring an execution time deviation of each subtask; determining importance degree data of each object according to a task knowledge graph; acquiring execution data of each object, and task execution analysis information fed back by other objects associated with an executed subtask; determining a compatibility feature of each object according to the execution data and the task execution analysis data; determining a representation vector of each object and a representation vector of each subtask according to the task knowledge graph; determining a target feature vector of each object according to the execution time deviation, the importance degree data, the compatibility feature, the representation vector of each object and the representation vector of each subtask; and obtaining a sorting result of a plurality of objects according to the target feature vectors. A sorting result, etc., can be written into a blockchain.

Description

Object sorting method, related equipment and medium

This application claims priority to a Chinese patent application filed with the China Patent Office on September 10, 2021 with application number 202111067106.X and titled "Object Sorting Method, Related Devices and Media", the entire contents of which are incorporated by reference in In this application.

technical field

The present application relates to the technical field of artificial intelligence, in particular to an object sorting method, related equipment and media.

Background technique

The inventor realizes that at present, the traditional sorting method usually sorts the multiple objects to be sorted by their own reference data. For example, sort the employees of the enterprise based on their daily work assessments to obtain a performance ranking, and then classify the employees according to the ranking results; another example, through the running tasks (tasks in the running state) in the cluster The scheduling situation sorts the running tasks to obtain the importance ranking, and then the priority of the running tasks can be determined according to the sorting results. However, only sorting is performed based on the object's own reference data, and the data is single, which easily leads to low accuracy of sorting results and low sorting efficiency.

Contents of the invention

Embodiments of the present application provide an object sorting method, related equipment, and media, which can improve the sorting efficiency and the accuracy of sorting results for objects.

On the one hand, the embodiment of the present application provides an object sorting method, the method comprising:

Obtain the execution time deviation of each subtask included in the task;

Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

determining the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.

On the one hand, an embodiment of the present application provides an object sorting device, the device comprising:

An acquisition module, configured to acquire the execution time deviation of each subtask included in the task;

A determining module, configured to determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes entities of each object and entities of each subtask;

The acquiring module is configured to acquire the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

The determination module is configured to determine, according to the execution data and the task execution analysis data, the compatibility feature of each object in the subtask executed by the object;

The determining module is further configured to determine the characterization vector of each object and the characterization vector of each subtask according to the task knowledge map;

The determination module is further configured to determine each The target feature vector of an object;

A sorting module, configured to sort the plurality of objects according to the target feature vector of each object to obtain a sorting result.

In one aspect, an embodiment of the present application provides an electronic device, the electronic device includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions, to execute the following method:

Obtain the execution time deviation of each subtask included in the task;

Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

determining the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.

On the one hand, an embodiment of the present application provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, the computer program includes program instructions, and when the program instructions are executed by a processor, they are used to perform the following methods:

Obtain the execution time deviation of each subtask included in the task;

Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

determining the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.

By implementing the method proposed in the embodiment of the present application, the data used for sorting is more comprehensive, and the efficiency of sorting objects and the accuracy of the sorting results obtained are improved.

Description of drawings

FIG. 1 is a schematic flow diagram of an object sorting method provided in an embodiment of the present application;

FIG. 2 is a schematic flow chart of an object sorting method provided in an embodiment of the present application;

FIG. 3 is a schematic diagram of a task knowledge map provided by the embodiment of the present application;

FIG. 4a is a schematic diagram of a scene for determining importance data provided by an embodiment of the present application;

FIG. 4b is a schematic diagram of a scenario for determining importance data provided by an embodiment of the present application;

FIG. 5 is a schematic structural diagram of an object sorting device provided in an embodiment of the present application;

FIG. 6 is a schematic structural diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

The object sorting method proposed in the embodiment of the present application is implemented in an electronic device, and the electronic device may be a terminal device or a server. Wherein, the terminal device may be a smart phone, a tablet computer, a notebook computer, a desktop computer, and the like. The server can be an independent server, or a server cluster or distributed system composed of multiple physical servers, or it can provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware Services, domain name services, security services, content delivery network (Content Delivery Network, CDN), and cloud servers for basic cloud computing services such as big data and artificial intelligence platforms, but are not limited to this. This application involves blockchain technology. Electronic devices can write related data into the blockchain, such as task knowledge graphs, execution time deviations of subtasks, or sorting results of multiple objects, so that electronic devices can Get the information you need, such as sorting results for multiple objects.

In some embodiments, the electronic device may execute the method for sorting objects according to actual business requirements, so as to improve the efficiency of sorting objects and the accuracy of sorting results. The technical solution of the present application can be applied to any object sorting scenario. For example, the technical solution of this application can be applied to the scenario of sorting the performance of employees, and the electronic device can generate a knowledge map (ie, a task knowledge map) based on the employee (such as an employee of an enterprise) and the sub-projects that the employee participates in, and combine the project knowledge map , employees, and the relevant execution information of the sub-projects that employees participate in (execution time deviation, execution data, etc.) to obtain the employee's target feature vector, and sort multiple employees based on the target feature vector to obtain the sorting result. As another example, the technical solution of the present application can be applied to the scenario of sorting the importance of running tasks in the cluster, and the electronic device can generate a task knowledge map according to the running tasks and the task events (such as obtaining business data) executed by the running tasks, and Combining the task knowledge graph, running tasks, and related execution information (execution time deviation, execution data, etc.) of the task events executed by the running tasks to obtain the target feature vector of the running task, and sort multiple running tasks based on the target feature vector, Get sorted results. Here, there is no limitation on the object sorting scenario of the application, that is, no limitation on the specific types of the involved objects and tasks (subtasks). For the sake of illustration, unless otherwise specified, the object sorting methods mentioned later have all taken the performance sorting scenario of employees as an example.

It can be understood that the above scenarios are only examples, and do not constitute limitations on the application scenarios of the technical solutions provided in the embodiments of the present application, and the technical solutions of the present application may also be applied to other scenarios. For example, those skilled in the art know that with the evolution of the system architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

Based on the above description, an embodiment of the present application proposes a method for sorting objects, which can be executed by the above-mentioned electronic device. As shown in Figure 1, the process of the object sorting method in the embodiment of the present application may include the following:

S101. Obtain the execution time deviation of each subtask included in the task.

Wherein, the task and each subtask may be any event associated with the object, and the object may be any object that needs to be sorted. There can be one or more tasks, and one or more subtasks included in a task. For example, if the object is an employee of an enterprise, the task can be a project of the enterprise, and the subtask can be a subproject in the project; if the object is a running task in a cluster, the task can be the cluster node requested by the running task at runtime (such as a business database), a subtask can be a task event executed when a running task requests a cluster node.

In a possible implementation, the electronic device can obtain the execution record corresponding to the task from the database, the execution record stores the actual execution time and expected execution time of each subtask included in the task, and based on the actual execution time of each subtask Execution time and expected execution time determine the execution time deviation of each subtask; or it can also obtain the actual execution time and expected execution time of each subtask from the task knowledge map, for example, find the entity of each subtask in the task knowledge map corresponding node, and obtain the stored information from the node properties. Therefore, the acquisition of the execution time deviation of each subtask included in the task by the electronic device may be specifically, acquiring the actual execution time of each subtask included in the task and the expected execution time of each subtask, and calculating the actual execution time of each subtask and the expected execution time of each subtask. The difference between the expected execution times of the tasks is to determine the execution time deviation of each subtask according to the difference between the actual execution time of each subtask and the expected execution time of each subtask. Specifically, the difference between the actual execution time of each subtask and the expected execution time of each subtask may be directly determined as the execution time deviation of each subtask. Wherein, the execution time deviation is used to measure the execution status of the subtask executed by the object, and the execution status of the subtask affects the sorting result of the object, that is, when sorting the objects, the execution time deviation of the subtask will be combined. An object can perform one or more subtasks, and a subtask can be performed by one or more objects. Optionally, it is possible to first determine the set of objects to be sorted (may be multiple), and determine the subtasks performed by each object in the set of objects to be sorted, and obtain the subtasks performed by each object Execution time skew.

S102. Determine the importance degree data of each of the multiple objects associated with the task according to the task knowledge graph.

Wherein, the above-mentioned task knowledge map includes the entity of each object and the entity of each subtask. Optionally, it may also include the entity of the task to which each subtask belongs, that is, there may be multiple object entities, multiple The entity of the task and the entity of multiple subtasks, as well as the connection relationship between objects, the connection relationship between multiple tasks, the connection relationship between multiple subtasks, the connection relationship between objects and tasks, the object There can be a connection relationship with a subtask, and there can be a connection relationship between a task and a subtask. For example, there is a connection relationship between each object and the subtasks executed by each object in each subtask. Optionally, there may also be a connection relationship between each subtask and the task it belongs to, and there may also be a connection relationship between tasks and The connection relationship between the associated objects (the objects that execute the subtasks included in the task can be used as multiple objects associated with the task).

In a possible implementation, the electronic device determines the importance data of each of the multiple objects associated with the task according to the task knowledge graph, specifically, by using the PageRank algorithm (a method used to classify nodes in a directed connection graph) Importance ranking algorithm) to obtain the importance of each entity in the task knowledge graph, and use the importance of each object's entity as the importance data of each object; or, use the PageRank algorithm to obtain The importance of the entity of each object in the task knowledge graph, and the importance of the entity of each subtask in the task knowledge graph, according to the importance of the entity of each object and the entity of the subtask performed by each object The importance determines the importance data of each object, for example, the ratio of the importance of the entity of each object to the importance of the entity of the subtask executed by each object may be used as the importance data of each object. The importance degree data can be used to evaluate the importance of the object, specifically, it can be used to evaluate the importance in the executed subtask, that is, the importance degree data of each object can be used to represent the importance of each object in the executed subtask Importance data in the task. The importance data of each object in the different subtasks executed can be the same or different. The importance of the object in the executed subtasks affects the sorting results of the objects. When sorting, the importance of the object in the subtask being performed is combined.

S103. Obtain the execution data of the subtasks executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtasks.

For example, if object 1 executes subtask 1, the execution data of object 1 on the executed subtask 1 is obtained, and the objects that execute subtask 1 also include object 2 and object 3, that is, for object 1, the executed subtask The other objects associated with task 1 are object 2 and object 3, so the task execution analysis information on object 1 fed back by other objects (part or all of objects in object 2 and object 3) associated with the executed subtask 1 is obtained.

In some embodiments, the electronic device can obtain the execution record corresponding to the task from the database, and the execution record stores the execution data of each object associated with each subtask included in the task when executing the subtask and other associated objects. Feedback task execution analysis information for this object, that is, for multiple objects executing subtasks, the execution record corresponding to the task contains the execution data generated by each object during the execution of the subtask, and multiple The task execution analysis information fed back by other objects in the object during the execution of the subtask.

In one embodiment, taking the object as an employee as an example, the execution data may represent the execution status of the object at different periods during the execution of subtasks, and the task execution analysis information may represent feedback feedback from other objects on the execution status. For example, the execution data can specifically be the work status records filled in by employee A during the execution of the subtask (i.e. the execution status in different periods), such as daily content, weekly report content, etc., and the task execution analysis information can be other information associated with the subtask. Employee B's work evaluation record of employee A's feedback (that is, the feedback feedback), such as the reply record of employee A's execution data, or the score of employee A's work participation, etc. In one embodiment, taking the running task as an example, the execution data may be the running log generated by the running task during the execution of the task event, and the task execution analysis information may be the data generated during the execution of other running tasks that execute the task event. The associated running log of the running task, etc. For example, the execution data may specifically be the operation exception record generated by the execution task A during the execution of the task event (that is, the operation log generated during the execution period), such as the operation error and operation suspension records of the execution task A, and the task execution analysis information may specifically be the execution of the task. The associated running exception record for the running task A generated by other running task B of the task event (that is, the associated running log for the running task A generated), such as the existence of other running tasks B caused by the abnormal running of the running task A Records of abnormal operation, etc. There is no limitation on specific types of execution data and task execution analysis information.

S104. Determine the compatibility feature of each object in the subtask executed by the object according to the execution data and the task execution analysis data.

In a possible implementation manner, the electronic device may determine the compatibility feature of each object in the subtask executed by the object according to the execution data and the task execution analysis data, which may include acquiring a first feature vector used to represent the execution data, and It is used to represent the second feature vector of the task execution analysis data, and the compatibility feature of the object in the executed sub-task is determined according to the first feature vector and the second feature vector. In one embodiment, the compatibility feature of the object in the executed subtask may be represented by a numerical value of 0-1. Among them, the compatibility feature can be used to indicate whether the execution data of the object in the executed subtask matches the task execution analysis data fed back by other objects to the object. The higher the matching degree, the higher the value of the compatibility feature, and vice versa , the lower the matching degree, the lower the value of the compatibility feature.

For example, taking the object as an employee, the execution data is the record of the employee’s work situation, and the task execution analysis data is the work evaluation record of the employee by other employees associated with the sub-project. The corresponding first eigenvector and The second eigenvector corresponding to the employee’s work evaluation record of other employees obtains the compatibility feature of the employee in this sub-item, and the compatibility feature indicates the consistency between the employee’s work record and the job evaluation record. If the compatibility The higher the sexual characteristics, the more consistent the work performance represented by the employee’s work situation and the work performance represented by other feedback job evaluations are in the participation of this sub-item, and vice versa. , employees with higher compatibility characteristics are more likely to be ranked high in performance.

S105. Determine the representation vector of each object and the representation vectors of each subtask according to the task knowledge map.

In a possible implementation, the electronic device determines the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph. Specifically, the entity of each object, the entity of each subtask, and the object and the relationship between the subtasks; the translation distance model is invoked to obtain the characterization vector of each object and the characterization vectors of each subtask according to the entity of each object, the entity of each subtask, and the relationship between the object and the subtask.

Among them, the translation distance model can be a TransE (Translation embeddings for modeling multi-relation data, multi-relational data embedding) model, so the electronic device can And use the translation distance model to obtain the representation vector of each object and the representation vector of each subtask. Specifically, it can be constructed according to the entity of each object, the entity of each subtask, and the relationship between the object and the subtask. tuples, the multiple triples are determined by the first entity, the second entity and the relationship between the first entity and the second entity, the first entity can be any object in each object, and the second entity can be is the subtask executed by the object, so the triplet can be (object, relationship, subtask), and the first entity, the second entity and the relationship in multiple triplets are mapped to the target vector space, and get The mapping vector of the first entity, the mapping vector of the second entity, and the mapping vector of the relationship, by constantly adjusting the mapping vector of the first entity, the mapping vector of the second entity, and the mapping vector of the relationship in the target vector space, so that The mapping vector of the first entity, the mapping vector of the second entity, and the mapping vector of the relationship in each triple group satisfy the preset relationship, and the mapping vector of the first entity when the preset relationship is satisfied is used as the representation vector of the object, The mapping vector of the second entity when the preset relationship is satisfied is used as the representation vector of the subtask of the object. Optionally, the preset relationship may be that the sum vector of the mapping vector of the first entity and the mapping vector of the relationship is similar to the mapping vector of the second entity, so each triplet (object, relationship, subtask) can be The relationship of is seen as the translation from the object entity to the subtask entity.

S106. Determine the target feature vector of each object according to the execution time deviation, the importance level data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.

In a possible implementation manner, the electronic device determines the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vector of each subtask. Specifically, it may be: Determine all subtasks performed by each object, and obtain the initial feature vector of each subtask based on the execution time deviation, importance data, compatibility features and the characterization vector of each subtask in all subtasks, and obtain the initial feature vector of each subtask according to the The representation vector and the initial feature vector of each subtask performed by the object are obtained to obtain the target feature vector of each object.

Wherein, the electronic device obtains the initial feature vector of each subtask based on the execution time deviation, the importance data, the compatibility feature and the characterization vector of each subtask in all the subtasks. Specifically, according to the execution time deviation of each subtask, The product of the importance data of the object in each subtask, the compatibility feature of the object in each subtask and the characterization vector of each subtask is obtained to obtain the initial feature vector of each subtask.

In some embodiments, according to the characterization vector of each object and the initial feature vector of each subtask performed by the object, the target feature vector of each object can be obtained specifically, the initial feature vector of each subtask performed by the object The summation is obtained to obtain the target initial feature vector of each object, and the target feature vector of each object is obtained according to the representation vector of each object and the target initial feature vector of each object. Wherein, according to the characterization vector of each object and the target initial eigenvector of each object, the target eigenvector of each object can be obtained by using the sum vector of the characterization vector and the target initial eigenvector as the target eigenvector, or by using the characterization vector The vector product with the target initial feature vector is used as the target feature vector.

S107, sort the multiple objects according to the target feature vector of each object, and obtain a sorting result.

In a possible implementation, the electronic device sorts the multiple objects according to the target feature vector of each object, and obtaining the sorting result may specifically include constructing the multiple objects into multiple object combinations, and according to the target feature vector of each object The vector determines a difference vector between each object combination in the multiple object combinations, sorts the multiple objects according to the difference vector between each object combination, and obtains a sorting result. Wherein, the object combination is any two objects in the plurality of objects, that is, every two objects in the plurality of objects form an object combination.

Optionally, the electronic device sorts the multiple objects according to the difference vector between each object combination, and obtaining the sorting result may be, according to the difference vector between each object combination, determining the sorting result of the objects in each object combination , and sort the multiple objects according to the sorting results of the objects in each object combination to obtain the sorting results of the multiple objects. For example, multiple objects include object 1, object 2, and object 3, and the constructed object combinations are combination 1 (object 1, object 2), combination 2 (object 2, object 3), combination 3 (object 1, object 3) , according to the target feature vector of each object in combination 1, determine the difference vector 1 between combination 1, and according to the difference vector 1, determine the sorting result of the objects in combination 1 as object 2, object 1, and according to each of combination 2 The target feature vector of the object determines the difference vector 2 between the combination 2, and according to the difference vector 2, it is determined that the object sorting result in the combination 2 is object 2 and object 3, and the combination is determined according to the target feature vector of each object in the combination 3 The difference vector 3 between 3, and according to the difference vector 3, it is determined that the object sorting result in the combination 3 is object 3 and object 1, so multiple objects are sorted according to the object sorting result in each object combination, and more The sorting result of objects is object 2, object 3, object 1.

In the embodiment of the present application, the electronic device can obtain the execution time deviation of each subtask included in the task, determine the importance data of each of the multiple objects associated with the task according to the task knowledge map, and obtain the subtask performed by each object on the object. The execution data of the task and the task execution analysis information of the object fed back by other objects associated with the executed sub-task, determine the compatibility characteristics of each object in the sub-task executed by the object according to the execution data and task execution analysis data, according to the task The knowledge map determines the characterization vector of each object and the characterization vector of each subtask, and determines the The target feature vector is used to sort multiple objects according to the target feature vector of each object to obtain the sorting result. By implementing the method proposed in the embodiment of the present application, a task knowledge graph including objects and tasks associated with objects can be generated, and multiple objects can be sorted in combination with the task knowledge graph, objects, and tasks, so that the data used in sorting can be compared. Comprehensive, and improve the efficiency of object-based sorting and the accuracy of the resulting sorting results.

Please refer to FIG. 2 . FIG. 2 is a schematic flowchart of a method for sorting objects provided by an embodiment of the present application, and the method may be executed by the electronic device mentioned above. As shown in Figure 2, the process of the object sorting method in the embodiment of the present application may include the following:

S201. Obtain the execution time deviation of each subtask included in the task.

In some embodiments, the electronic device may determine the execution time deviation according to the actual execution time and expected execution time of the subtasks. Specifically, the actual execution time of each subtask included in the task and the expected execution time of each subtask may be obtained, and the calculation The difference between the actual execution time of each subtask and the expected execution time of each subtask, according to the difference between the actual execution time of each subtask and the expected execution time of each subtask, determine the execution time of each subtask deviation. Wherein, the difference between the actual execution time of each subtask and the expected execution time of each subtask may be used as the execution time deviation of each subtask to be determined. For example, the actual execution time of subtask j is DT _j , and the expected execution time is T _j , so the execution time deviation of subtask j may be diff _j =DT _j −T _j .

S202. Determine the importance degree data of each of the multiple objects associated with the task according to the task knowledge graph.

In a possible implementation, taking the object as an employee as an example, the electronic device may construct a task knowledge map by extracting multiple enterprise employees from enterprise information, and extracting items (tasks) associated with the multiple enterprise employees , and extract the sub-projects (sub-tasks) to which the project belongs, according to the relationship between employees, the relationship between employees and projects, the relationship between employees and sub-projects, the relationship between projects and sub-projects, and the relationship between sub-projects The relationship between constructing task knowledge graph, for example, the constructed task knowledge graph can be shown in Figure 3; among them, the relationship between employees can be used to indicate the superior-subordinate relationship between employees, and the relationship between sub-projects can be used for Indicates dependencies of multiple subprojects under the same project. For example, a project includes subproject A, subproject B, and subproject C. If subproject A needs to be completed before subproject B can be completed, subproject A is the predecessor subproject of subproject B, that is, subproject A is subproject B. Dependent subprojects, and if subproject B can be completed before subproject C is completed, then subproject B is the predecessor subproject of subproject C, and all the predecessor subprojects of subproject C are subproject A and subproject B, namely Subproject A and Subproject B are subprojects that Subproject C depends on.

Optionally, the task knowledge map can also be integrated into the organizational structure, such as the department to which the employee belongs. Optionally, the electronic device can obtain real-time project information or real-time employee information from enterprise information, and update the task knowledge map in real time, so that the timeliness of the employee performance ranking obtained according to the task knowledge map can be improved. In the future, the electronic device can dynamically adjust the performance ranking of employees by combining the task knowledge map that integrates the relevant information of enterprise employees and project-related information and the relevant work status of the sub-projects that employees participate in, which can improve the efficiency and accuracy of employee ranking.

In a possible implementation manner, the electronic device determines the importance data of each of the multiple objects associated with the task according to the task knowledge map. Specifically, the PageRank algorithm is used to calculate the value of each entity in the task knowledge map. , and use the importance of the entity of each object as the importance data. Wherein, using the PageRank algorithm to calculate the importance of each entity in the task knowledge graph can be specifically constructed by constructing an adjacency matrix containing the connection relationship of each entity according to the task knowledge graph, and the adjacency matrix represents the relationship between each entity. Whether there is a connection relationship between them, and the direction of the connection relationship, set the initial importance of each entity as 1, then generate a transition matrix containing each entity according to the adjacency matrix, that is, the value of each row in the adjacency matrix can be made The normalization process obtains a transfer matrix, the sum of the values of each row in the transfer matrix is 1, and a system of equations for each entity is constructed according to the transfer matrix, which is a plurality of functions about entities, which can be solved through the system of equations Get the importance of each entity.

For example, as shown in Figure 4a-Figure 4b, Figure 4a-Figure 4b is a schematic diagram of a scene for determining the importance level data provided by the embodiment of the present application, let Figure 4a be a task knowledge map, and the entities contained are ①-⑥, according to The task knowledge map constructs an adjacency matrix that contains the connection relationship of each entity, as shown in Figure 4b(1). The connection relationship between entities is expressed in the form of an adjacency matrix. For example, matrix [①][⑥]=1 means from entity ① to Entity ⑥ has a connection relationship and is directed from entity ① to entity ⑥. The initial importance of each entity is set to 1, so the transition matrix generated according to the adjacency matrix can be seen in Figure 4b(2), where the value of each row in the transition matrix is The sum=1, the equation system constructed according to the transfer matrix can be seen in Fig. 4b(3), the importance degree of each entity can be obtained by solving the equation system, and then the importance degree data of each object can be obtained.

S203. Obtain the execution data of the subtasks executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtasks. Wherein, for specific implementation manners of steps S202-S203, reference may be made to relevant descriptions of steps S102-S103, which will not be repeated here.

S204. Analyze data according to execution data and task execution.

In a possible implementation, the electronic device determines the compatibility characteristics of each object in the subtasks executed by the object according to the execution data and the task execution analysis data. Specifically, the execution data may be vectorized to obtain the corresponding The first eigenvector of the task execution analysis data is vectorized to obtain the second eigenvector corresponding to the task execution analysis data, and the first eigenvector and the second eigenvector are input into the classification model to obtain the performance of each object in the object execution Compatibility traits in subtasks. Among them, vectorizing the execution data can be inputting the execution data into the BERT (Bidirectional Encoder Representation from Transformers, bidirectional encoder of the modeler) model to obtain the first feature vector, and vectorizing the task execution analysis data can be The task execution analysis data is input into the BERT model to obtain the second feature vector.

Optionally, the classification model can be a sigmoid neural network model, so the compatibility feature of object i in subtask j can be:

Among them, σ is a sigmoid neural network model, M is a model parameter in the model, specifically M can be a bilinear parameter matrix, rV _i ^j is the first eigenvector corresponding to the execution data of object i in subtask j, dV _i ^j is the second feature vector corresponding to the task execution analysis data fed back by other objects to object i in subtask j.

S205. Determine the representation vector of each object and the representation vectors of each subtask according to the task knowledge graph.

In a possible implementation manner, the electronic device determines the characterization vector of each object and the characterization vector of each subtask according to the task knowledge graph. Specifically, the entity of each object and the entity of each subtask are obtained from the task knowledge graph, Perform vector representation on the entity of each object and the entity of each subtask, obtain the word vector of each object and the word vector of each subtask, and perform weighted processing on the word vector of each object and the word vector of each subtask, Obtain the relationship mapping vector of each object and the relationship mapping vector of each subtask, determine the relationship mapping vector of each object as the representation vector of each object, and determine the relationship mapping vector of each subtask as the representation of each subtask vector. Wherein, the electronic device performs vector representation on the entity of each object and the entity of each subtask respectively, and a vector dictionary is established in advance, and the corresponding relationship between word vectors and entities is stored in the dictionary, and the meanings of the entities in the vector dictionary are similar. The distance between the word vectors of the entities is also similar, so the entities of each object and the entities of each subtask can be represented by vectors based on the vector dictionary, or the word2vec tool can be used to build a word vector model, and the word vector The model is trained so that the trained word vector model can output the word vector corresponding to each entity, and the closer the word meaning is to the entity corresponding to the closer the word vector distance, so the entity of each object and the entity of each subtask can be separated Input to the trained word vector model, the word vector model performs vector representation and outputs the word vector of each object and the word vector of each subtask; the word vector distance can point to the Hamming distance, Euclidean distance, etc. between quantities.

In some embodiments, the electronic device performs weighting processing on the word vector of each object and the word vector of each subtask, specifically, determining the relationship between the object and the subtask from the task knowledge map, and obtaining the object and subtask The relational mapping matrix corresponding to the relationship among them is used to weight the word vector of each object and the word vector of each subtask respectively. Wherein, the relationship between each object and the executed subtask is the same, so the relationship mapping matrix used for weighting the word vector of each object and the word vector of each subtask is also the same. Weighting the word vectors of each object and the word vectors of each subtask through the relationship mapping matrix corresponding to the relationship can be understood as mapping each object to the relationship space where the relationship is located, and mapping each subtask to the relationship where the relationship is located. relational space.

Optionally, the electronic device may obtain a relationship mapping matrix corresponding to the relationship through the translation distance model. Optionally, the translation distance model can be a TransR (Learning Entity and Relation Embeddings for Knowledge Graph Completion, entities and relations are embedded separately) model, so the specific way for the electronic device to obtain the relationship mapping matrix can be, for the sample object and the sample subtask The sample relationship between samples is expressed as a vector, and the sample word vector of the sample relationship is obtained, and the sample relationship mapping matrix is constructed, and the sample word vector of the sample object, the sample subtask and the sample word vector are weighted by the sample relationship mapping matrix, and the sample The sample representation vector of the object and the sample representation vector of the sample subtask construct the objective function, and use the sample representation vector of the sample object, the sample representation vector of the sample subtask and the sample word vector of the sample relationship to map the sample relationship mapping matrix according to the objective function Training to obtain the relationship mapping matrix corresponding to the relationship between the above objects and subtasks; the sample object and the object in the task knowledge map can be the same or different, and the sample subtasks and the subtasks in the task knowledge map can be the same or different . That is, the objective function can be:

Among them, h represents the sample word vector of the sample object; r represents the sample word vector of the sample relationship between the sample object and the sample subtask; t represents the sample word vector of the sample subtask; h _r = hM _r represents the sample representation of the sample object vector; M _r represents the sample relationship mapping matrix, and the process of using this M _r to weight h to obtain h _r is to map the sample word vector of the sample object to the sample relation space corresponding to the sample relation; t _r =tM _r represents the sample The sample representation vector of the subtask; the process of using this M _r to weight t to obtain t _r is to map the sample word vector of the sample subtask to the sample relational space corresponding to the sample relation; therefore, the purpose of training is to make the representation vector of the object The sum vector of the word vector of the relationship is approximately equal to the representation vector of the subtask.

S206. Determine the target feature vector of each object according to the execution time deviation, the importance level data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.

In a possible implementation, subtasks may have predecessor subtasks, that is, subtasks with dependencies. When the execution time deviation of dependent subtasks is large, the execution time deviation of the current subtask may be relatively small. Large, so the impact of dependent subtasks on the execution time deviation of the current subtask can be offset by accumulating execution time deviations. That is, the electronic device determines the target feature vector of each object according to the execution time deviation, importance data, compatibility features, characterization vectors of each object, and characterization vectors of each subtask. Task, the sum of the execution time deviation of each subtask and the execution time deviation of the subtasks that each subtask depends on is determined as the cumulative execution time deviation of each subtask, according to the cumulative execution time deviation, importance data, and compatibility characteristics , the characterization vector of each object and the characterization vectors of each subtask, and determine the target feature vector of each object. Wherein, the electronic device determines the subtasks on which each subtask depends may obtain the execution record corresponding to the task from the database, and the execution record stores the execution links of the subtasks, which means the relationship between the subtasks under one task; Alternatively, the subtasks that each subtask depends on can be obtained from the task knowledge graph according to the connection relationship between each subtask. Therefore, the cumulative execution time deviation of each subtask can be:

Among them, TB _j is the cumulative execution time deviation of subtask j, and diff _k represents the execution time deviation between the subtask and the subtasks that the subtask depends on.

For example, suppose a task includes subtask 1, subtask 2, subtask 3, subtask 3 is a subtask that subtask 2 depends on, subtask 3 and subtask 2 are subtasks that subtask 1 depends on, if the execution of subtask 1 The time deviation is diff ₁ , the execution time deviation of subtask 2 is diff ₂ , and the execution time deviation of subtask 3 is diff ₃ , so the cumulative execution time deviation of subtask 1 is TB ₁ =diff ₁ +diff ₂ +diff ₃ , The cumulative execution time deviation of subtask 2 is TB ₂ =diff ₂ +diff ₃ , and the cumulative execution time deviation of subtask 3 is TB ₃ =diff ₃ .

In a possible implementation, the process and principle of determining the target feature vector of each object are the same, and here only the determination of the target feature vector of an object is taken as an example for illustration; Compatibility features, characterization vectors of each object and characterization vectors of each subtask, determining the target feature vector of each object can specifically be to determine all target subtasks performed by the target object, according to the target object in all target subtasks The cumulative execution time deviation, important program data and compatibility features of each subtask in , the characterization vector of each target subtask, determine the initial feature vector of each target subtask, and convert the initial feature vector of each target subtask And the representation vector of the target object is input into the preset neural network model to obtain the target feature vector of the target object; the target object is any object in each object.

Optionally, the preset neural network model can be a transformer model, so the target feature vector eC _i of the target object i can be:

Among them, eV _i represents the representation vector of target object i; j ∈ J represents all target subtasks; TB _j represents the cumulative execution time deviation of target subtask j; pV _i represents the representation vector of target subtask j;

Indicates the importance data of the target object i in the target subtask j, and the importance data of the target object i in different target subtasks may be the same or different;

Indicates the compatibility feature of target object i in target subtask j.

S207. Construct multiple objects into multiple object combinations. Wherein, for the specific implementation manner of step S207, reference may be made to the related description of step S107.

S208. Sorting the multiple objects according to the target feature vector of each object and each of the multiple object combinations, to obtain a sorting result.

In a possible implementation manner, the electronic device sorts the multiple objects according to the target feature vector of each object and each object combination in the multiple object combinations. Specifically, it may be to determine the multiple objects according to the target feature vector of each object The difference vector between each combination of objects in the combination, sorting multiple objects by the difference vector between each combination of objects. Wherein, sorting the multiple objects according to the difference vector between each object combination may be specifically, according to the difference vector between each object combination, determining the object sorting result in each object combination, and according to each object combination Sort multiple objects within the object sort result. Wherein, according to the difference vector between each object combination, determining the object sorting result in each object combination may specifically include inputting the difference vector between each object combination into the classification model, and determining each object according to the output result of the classification model. The result of sorting the objects within the object group. The electronic device can obtain a sorting result including multiple objects according to the sorting result of objects in each object combination.

Optionally, the classification model may be a sigmoid neural network model. It can be understood that the classification model used to determine the object sorting result in each object combination here may not be the same model as the above-mentioned classification model used to determine the compatibility feature of the object in the executed subtask.

Therefore, assuming the target feature vector eC _u of object u and the target feature vector eC _v of object v in the object combination, the difference vector is W _u,v = σ(eC _u -eC _v ), if W _u,v is input into the classification model, if the obtained output value is greater than 0.5, then the object sorting result in the object combination indicates that the ranking r _u of object u is greater than the ranking r v of object _v , that is, r _u >r _v .

In the embodiment of the present application, the electronic device can obtain the execution time deviation of each subtask included in the task, determine the importance data of each of the multiple objects associated with the task according to the task knowledge map, and obtain the subtask performed by each object on the object. The execution data of the task and the task execution analysis information of the object fed back by other objects associated with the executed sub-task, determine the compatibility characteristics of each object in the sub-task executed by the object according to the execution data and task execution analysis data, according to the task The knowledge map determines the characterization vector of each object and the characterization vector of each subtask, and determines the The target feature vector constructs multiple objects into multiple object combinations, sorts the multiple objects according to the target feature vector of each object and each object combination in the multiple object combinations, and obtains a sorting result. By implementing the method proposed in the embodiment of the present application, a task knowledge graph including objects and tasks associated with objects can be generated, and multiple objects can be sorted in combination with the task knowledge graph, objects, and tasks, so that the data used in sorting can be compared. Comprehensive, and improve the efficiency of sorting objects and the accuracy of the resulting sorting results.

Please refer to FIG. 5 , which is a schematic structural diagram of an object sorting device provided in the present application. It should be noted that the object sorting device shown in FIG. 5 is used to execute the method of the embodiment shown in FIG. 1 and FIG. 2 of the present application. For the convenience of description, only the parts related to the embodiment of the present application are shown. The specific technology The details are not disclosed, and reference is made to the embodiment shown in Fig. 1 and Fig. 2 of the present application. The object sorting apparatus 500 may include: an acquisition module 501 , a determination module 502 , and a sorting module 503 . in:

An acquisition module 501, configured to acquire the execution time deviation of each subtask included in the task;

A determining module 502, configured to determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entities of each subtask;

The obtaining module 501 is configured to obtain the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

The determination module 502 is configured to determine, according to the execution data and the task execution analysis data, the compatibility feature of each object in the subtask executed by the object;

The determination module 502 is further configured to determine the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

The determining module 502 is further configured to determine the target feature vector for each object;

A sorting module 503, configured to sort the plurality of objects according to the target feature vector of each object to obtain a sorting result.

In a possible implementation manner, the determining module 502 is configured to perform the following tasks according to the execution time deviation, the importance data, the compatibility feature, the characterization vector of each object, and the subtasks. The characterization vector, when determining the target feature vector of each object, is specifically used for:

determining the subtasks on which each subtask depends;

determining the sum of the execution time deviations of the subtasks and the execution time deviations of the subtasks on which the subtasks depend as the cumulative execution time deviation of the subtasks;

The target feature vector of each object is determined according to the cumulative execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.

In a possible implementation manner, when the obtaining module 501 is used to obtain the execution time deviation of each subtask included in the task, it is specifically used to:

Obtaining the actual execution time of each subtask included in the task and the expected execution time of each subtask;

calculating the difference between the actual execution time of each subtask and the expected execution time of each subtask;

The difference between the actual execution time of each subtask and the expected execution time of each subtask is determined as the execution time deviation of each subtask.

In a possible implementation manner, when the determination module 502 is used to determine the compatibility feature of each object in the subtask executed by the object according to the execution data and the task execution analysis data, Specifically for:

performing vectorization processing on the execution data to obtain a first feature vector corresponding to the execution data;

performing vectorization processing on the task execution analysis data to obtain a second feature vector corresponding to the task execution analysis data;

Inputting the first feature vector and the second feature vector into a classification model to obtain the compatibility feature of each object in the subtask performed by the object.

In a possible implementation manner, when the determining module 502 is used to determine the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph, it is specifically used to:

Obtain the entity of each object and the entity of each subtask from the task knowledge graph;

performing vector representation on the entity of each object and the entity of each subtask respectively, to obtain the word vector of each object and the word vector of each subtask;

performing weighting processing on the word vector of each object and the word vector of each subtask to obtain the relationship mapping vector of each object and the relationship mapping vector of each subtask;

The relationship mapping vector of each object is determined as the characterization vector of each object, and the relationship mapping vector of each subtask is determined as a characterization vector of each subtask.

In a possible implementation manner, the determining module 502 is specifically configured to:

determining a relationship between the object and the subtask from the task knowledge graph;

Obtaining a relationship mapping matrix corresponding to the relationship between the object and the subtask;

The word vector of each object and the word vector of each subtask are weighted by using the relationship mapping matrix.

In a possible implementation manner, when the sorting module 503 is used to sort the multiple objects according to the target feature vector of each object to obtain a sorting result, it is specifically configured to:

constructing the plurality of objects as a plurality of object combinations;

determining a difference vector between each of the plurality of object combinations according to the target feature vector of each of the objects;

The plurality of objects are sorted according to the difference vector between each combination of objects to obtain a sorting result.

In the embodiment of the present application, the acquisition module acquires the execution time deviation of each subtask included in the task; the determination module determines the importance data of each of the multiple objects associated with the task according to the task knowledge graph, and the task knowledge graph includes each object The entity of the entity and the entity of each subtask; the acquisition module acquires the execution data of the subtask executed by each object on the object and the task execution analysis information of the object fed back by other objects associated with the executed subtask; the determination module according to the execution data and task execution analysis data to determine the compatibility characteristics of each object in the subtasks executed by the object; the determination module determines the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph; the determination module determines the characterization vector according to the execution time deviation, Importance data, compatibility features, characterization vectors of each object and characterization vectors of each subtask, determine the target feature vector of each object; the sorting module sorts multiple objects according to the target feature vector of each object, and obtains Sort results. By implementing the device proposed in the embodiment of the present application, a task knowledge graph including objects and tasks associated with the objects can be generated, and multiple objects can be sorted in combination with the task knowledge graph, objects, and tasks, so that the data used in sorting can be compared. Comprehensive, and improve the efficiency of object-based sorting and the accuracy of the resulting sorting results.

Each functional module in each embodiment of the present application may be integrated into one module, each module may exist separately physically, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software function modules, which is not limited in this application.

Please refer to FIG. 6 , which is a schematic structural diagram of an electronic device provided in an embodiment of the present application. As shown in FIG. 6 , the electronic device 600 includes: at least one processor 601 and a memory 602 . Optionally, the electronic device may also include a network interface. Wherein, the processor 601, the memory 602 and the network interface can exchange data, the network interface is controlled by the processor 601 for sending and receiving messages, and the memory 602 is used for storing computer programs, and the computer programs include program instructions, The processor 601 is used to execute program instructions stored in the memory 602 . Wherein, the processor 601 is configured to call the program instruction to execute the above method.

The memory 602 may include a volatile memory (volatile memory), such as a random-access memory (random-access memory, RAM); the memory 602 may also include a non-volatile memory (non-volatile memory), such as a flash memory (flash memory), solid-state drive (solid-state drive, SSD) etc.; Described memory 602 can also comprise the combination of above-mentioned types of memory.

The processor 601 may be a central processing unit (central processing unit, CPU). In one embodiment, the processor 601 may also be a Graphics Processing Unit (GPU). The processor 601 may also be a combination of a CPU and a GPU.

In a possible implementation manner, the memory 602 is used to store program instructions, and the processor 601 can invoke the program instructions to perform the following steps:

Obtain the execution time deviation of each subtask included in the task;

Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

determining the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.

In a possible implementation manner, the processor 601 is configured to: The characterization vector, when determining the target feature vector of each object, is specifically used for:

determining the subtasks on which each subtask depends;

determining the sum of the execution time deviations of the subtasks and the execution time deviations of the subtasks on which the subtasks depend as the cumulative execution time deviation of the subtasks;

The target feature vector of each object is determined according to the cumulative execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.

In a possible implementation manner, when the processor 601 is used to obtain the execution time deviation of each subtask included in the task, it is specifically used to:

Obtaining the actual execution time of each subtask included in the task and the expected execution time of each subtask;

calculating the difference between the actual execution time of each subtask and the expected execution time of each subtask;

The difference between the actual execution time of each subtask and the expected execution time of each subtask is determined as the execution time deviation of each subtask.

In a possible implementation manner, when the processor 601 is used to determine the compatibility feature of each object in the subtask executed by the object according to the execution data and the task execution analysis data, Specifically for:

performing vectorization processing on the execution data to obtain a first feature vector corresponding to the execution data;

performing vectorization processing on the task execution analysis data to obtain a second feature vector corresponding to the task execution analysis data;

Inputting the first feature vector and the second feature vector into a classification model to obtain the compatibility feature of each object in the subtask performed by the object.

In a possible implementation manner, when the processor 601 is used to determine the characterization vector of each object and the characterization vector of each subtask according to the task knowledge graph, it is specifically configured to:

Obtain the entity of each object and the entity of each subtask from the task knowledge graph;

performing vector representation on the entity of each object and the entity of each subtask respectively, to obtain the word vector of each object and the word vector of each subtask;

performing weighting processing on the word vector of each object and the word vector of each subtask to obtain the relationship mapping vector of each object and the relationship mapping vector of each subtask;

The relationship mapping vector of each object is determined as the characterization vector of each object, and the relationship mapping vector of each subtask is determined as a characterization vector of each subtask.

In a possible implementation manner, when the processor 601 performs weight processing on the word vector of each object and the word vector of each subtask, it is specifically configured to:

determining a relationship between the object and the subtask from the task knowledge graph;

Obtaining a relationship mapping matrix corresponding to the relationship between the object and the subtask;

The word vector of each object and the word vector of each subtask are weighted by using the relationship mapping matrix.

In a possible implementation manner, when the processor 601 is configured to sort the multiple objects according to the target feature vector of each object to obtain a sorting result, it is specifically configured to:

constructing the plurality of objects as a plurality of object combinations;

determining a difference vector between each of the plurality of object combinations according to the target feature vector of each of the objects;

The plurality of objects are sorted according to the difference vector between each combination of objects to obtain a sorting result.

In a specific implementation, the device, processor 601, memory 602, etc. described in the embodiments of this application can execute the implementation methods described in the above method embodiments, and can also execute the implementation methods described in the embodiments of this application, which will not be repeated here repeat.

An embodiment of the present application also provides a computer (readable) storage medium, the computer storage medium stores a computer program, the computer program includes program instructions, and when the program instructions are executed by a processor, the processor Part or all of the steps performed in the foregoing method embodiments may be performed. Optionally, the computer storage medium may be volatile or non-volatile. The computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application program required by at least one function, etc.; Use the created data etc.

Among them, the blockchain referred to in this application is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain (Blockchain), essentially a decentralized database, is a series of data blocks associated with each other using cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

The "plurality" mentioned herein means two or more. "And/or" describes the association relationship of associated objects, indicating that there may be three types of relationships, for example, A and/or B may indicate: A exists alone, A and B exist simultaneously, and B exists independently. The character "/" generally indicates that the contextual objects are an "or" relationship.

Those of ordinary skill in the art can understand that realizing all or part of the processes in the methods of the above embodiments can be completed by instructing related hardware through a computer program. The program can be stored in a computer storage medium, and the computer storage medium can be As for the computer-readable storage medium, when the program is executed, it may include the processes of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM), etc.

What is disclosed above is only part of the embodiments of the application, and of course it cannot limit the scope of rights of the application. Those of ordinary skill in the art can understand the whole or part of the process of realizing the above embodiments, and make according to the claims of the application The equivalent changes still belong to the scope covered by this application.

Claims (20)

1. A method for sorting objects, wherein the method includes:

   Obtain the execution time deviation of each subtask included in the task;

   Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

   Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

   determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

   determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

   determining the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

   sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.
2. The method according to claim 1, wherein the said execution time deviation, said importance level data, said compatibility feature, said each object's characterization vector and said each subtask's characterization vector , to determine the target feature vector of each object, including:

   determining the subtasks on which each subtask depends;

   determining the sum of the execution time deviations of the subtasks and the execution time deviations of the subtasks on which the subtasks depend as the cumulative execution time deviation of the subtasks;

   The target feature vector of each object is determined according to the cumulative execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.
3. The method according to claim 1, wherein the execution time deviation of each subtask included in the acquisition task comprises:

   Obtaining the actual execution time of each subtask included in the task and the expected execution time of each subtask;

   calculating the difference between the actual execution time of each subtask and the expected execution time of each subtask;

   The difference between the actual execution time of each subtask and the expected execution time of each subtask is determined as the execution time deviation of each subtask.
4. The method according to claim 1, wherein said determining the compatibility characteristics of each object in the subtask executed by the object according to the execution data and the task execution analysis data comprises:

   performing vectorization processing on the execution data to obtain a first feature vector corresponding to the execution data;

   performing vectorization processing on the task execution analysis data to obtain a second feature vector corresponding to the task execution analysis data;

   Inputting the first feature vector and the second feature vector into a classification model to obtain the compatibility feature of each object in the subtask performed by the object.
5. The method according to claim 1, wherein said determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph comprises:

   Obtain the entity of each object and the entity of each subtask from the task knowledge graph;

   performing vector representation on the entity of each object and the entity of each subtask respectively, to obtain the word vector of each object and the word vector of each subtask;

   performing weighting processing on the word vector of each object and the word vector of each subtask to obtain the relationship mapping vector of each object and the relationship mapping vector of each subtask;

   The relationship mapping vector of each object is determined as the characterization vector of each object, and the relationship mapping vector of each subtask is determined as a characterization vector of each subtask.
6. The method according to claim 5, wherein, the word vector of each object and the word vector of each subtask are weighted, comprising:

   determining a relationship between the object and the subtask from the task knowledge graph;

   Obtaining a relationship mapping matrix corresponding to the relationship between the object and the subtask;

   The word vectors of each object and the word vectors of the subtasks are respectively weighted by using the relationship mapping matrix.
7. The method according to claim 1, wherein said sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result comprises:

   constructing the plurality of objects as a plurality of object combinations;

   determining a difference vector between each of the plurality of object combinations according to the target feature vector of each of the objects;

   The plurality of objects are sorted according to the difference vector between each combination of objects to obtain a sorting result.
8. An object sorting device, wherein the device comprises:

   An acquisition module, configured to acquire the execution time deviation of each subtask included in the task;

   A determining module, configured to determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes entities of each object and entities of each subtask;

   The acquiring module is configured to acquire the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

   The determining module is configured to determine, according to the execution data and the task execution analysis data, the compatibility feature of each object in the subtask executed by the object;

   The determining module is further configured to determine the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

   The determination module is further configured to determine each The target feature vector of an object;

   A sorting module, configured to sort the plurality of objects according to the target feature vector of each object to obtain a sorting result.
9. An electronic device, which includes a processor and a memory, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to invoke the program instructions to perform the following method:

   Obtain the execution time deviation of each subtask included in the task;

   Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

   Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

   determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

   determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

   determining the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

   sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.
10. The electronic device according to claim 9 , wherein performing the execution according to the execution time deviation, the importance degree data, the compatibility feature, the characterization vector of each object and the respective subtasks Characterizing vectors, determining the target feature vectors of each object, comprising:

    determining the subtasks on which each subtask depends;

    determining the sum of the execution time deviations of the subtasks and the execution time deviations of the subtasks on which the subtasks depend as the cumulative execution time deviation of the subtasks;

    The target feature vector of each object is determined according to the cumulative execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.
11. The electronic device according to claim 9, wherein the execution time deviation of each subtask included in the acquisition task comprises:

    Obtaining the actual execution time of each subtask included in the task and the expected execution time of each subtask;

    calculating the difference between the actual execution time of each subtask and the expected execution time of each subtask;

    The difference between the actual execution time of each subtask and the expected execution time of each subtask is determined as the execution time deviation of each subtask.
12. The electronic device according to claim 9, wherein performing the determining the compatibility feature of each object in the subtask executed by the object according to the execution data and the task execution analysis data comprises:

    performing vectorization processing on the execution data to obtain a first feature vector corresponding to the execution data;

    performing vectorization processing on the task execution analysis data to obtain a second feature vector corresponding to the task execution analysis data;

    Inputting the first feature vector and the second feature vector into a classification model to obtain the compatibility feature of each object in the subtask performed by the object.
13. The electronic device according to claim 9, wherein performing the determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph comprises:

    Obtain the entity of each object and the entity of each subtask from the task knowledge graph;

    performing vector representation on the entity of each object and the entity of each subtask respectively, to obtain the word vector of each object and the word vector of each subtask;

    performing weighting processing on the word vector of each object and the word vector of each subtask to obtain the relationship mapping vector of each object and the relationship mapping vector of each subtask;

    The relationship mapping vector of each object is determined as the characterization vector of each object, and the relationship mapping vector of each subtask is determined as a characterization vector of each subtask.
14. The electronic device according to claim 9, wherein performing the sorting of the plurality of objects according to the target feature vector of each object to obtain a sorting result includes:

    constructing the plurality of objects as a plurality of object combinations;

    determining a difference vector between each of the plurality of object combinations according to the target feature vector of each of the objects;

    The plurality of objects are sorted according to the difference vector between each combination of objects to obtain a sorting result.
15. A computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, the computer program includes program instructions, and when executed by a processor, the program instructions cause the processor to perform the following method:

    Obtain the execution time deviation of each subtask included in the task;

    Determine the importance data of each of the plurality of objects associated with the task according to the task knowledge graph; the task knowledge graph includes the entity of each object and the entity of each subtask;

    Obtaining the execution data of the subtask executed by each object on the object and the task execution analysis information on the object fed back by other objects associated with the executed subtask;

    determining compatibility features of each object in subtasks executed by the object according to the execution data and the task execution analysis data;

    determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph;

    Determine the target feature vector of each object according to the execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask;

    sorting the plurality of objects according to the target feature vector of each object to obtain a sorting result.
16. The computer-readable storage medium according to claim 15, wherein performing said execution based on said execution time offset, said importance data, said compatibility features, said each object's characterization vector and said each The characterization vector of the subtask determines the target feature vector of each object, including:

    determining the subtasks on which each subtask depends;

    determining the sum of the execution time deviations of the subtasks and the execution time deviations of the subtasks on which the subtasks depend as the cumulative execution time deviation of the subtasks;

    The target feature vector of each object is determined according to the cumulative execution time deviation, the importance data, the compatibility feature, the feature vector of each object, and the feature vectors of each subtask.
17. The computer-readable storage medium according to claim 15, wherein the execution time deviation of each subtask included in the acquisition task comprises:

    Obtaining the actual execution time of each subtask included in the task and the expected execution time of each subtask;

    calculating the difference between the actual execution time of each subtask and the expected execution time of each subtask;

    The difference between the actual execution time of each subtask and the expected execution time of each subtask is determined as the execution time deviation of each subtask.
18. The computer-readable storage medium according to claim 15, wherein performing said determining compatibility characteristics of said each object in subtasks executed by said object based on said execution data and said task execution analysis data ,include:

    performing vectorization processing on the execution data to obtain a first feature vector corresponding to the execution data;

    performing vectorization processing on the task execution analysis data to obtain a second feature vector corresponding to the task execution analysis data;

    Inputting the first feature vector and the second feature vector into a classification model to obtain the compatibility feature of each object in the subtask performed by the object.
19. The computer-readable storage medium according to claim 15, wherein performing the determining the characterization vector of each object and the characterization vectors of each subtask according to the task knowledge graph comprises:

    Obtain the entity of each object and the entity of each subtask from the task knowledge graph;

    performing vector representation on the entity of each object and the entity of each subtask respectively, to obtain the word vector of each object and the word vector of each subtask;

    performing weighting processing on the word vector of each object and the word vector of each subtask to obtain the relationship mapping vector of each object and the relationship mapping vector of each subtask;

    The relationship mapping vector of each object is determined as the characterization vector of each object, and the relationship mapping vector of each subtask is determined as a characterization vector of each subtask.
20. The computer-readable storage medium according to claim 15, wherein performing the sorting of the plurality of objects according to the target feature vector of each object to obtain a sorting result includes:

    constructing the plurality of objects as a plurality of object combinations;

    determining a difference vector between each of the plurality of object combinations according to the target feature vector of each of the objects;

    The plurality of objects are sorted according to the difference vector between each combination of objects to obtain a sorting result.

Images