Task planning method, system, electronic device, storage medium and program product

Abstract

The embodiment of the application provides a kind of task planning method, system, electronic equipment, storage medium and program product, it is related to artificial intelligence technical field, method includes: receiving object creates the planning request of task, planning request includes the task description information of object creates task;Intelligent planning module is called, according to task description information and the auxiliary information associated with object creates task, the target creation parameter of object creates task is determined, auxiliary information is used to represent the constraint condition of object creates task;Intelligent screening module is called, at least one first resource matched with target creation parameter is screened in resource pool, and the first resource is used to create the to-be-created object corresponding to the object created by object creation task;According to target creation parameter and at least one first resource, the task planning information of object creates task is generated.The technical scheme of the embodiment of the application improves the accuracy of task planning.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a task planning method, system, electronic device, storage medium, and program product. Background Technology

[0002] With the rapid development of internet technology, the number of digital objects such as online service platforms and knowledge bases has also increased. Before creating a digital object, the creation task is usually planned first. Currently, the planning of creation tasks mainly relies on manual analysis of creation requirements and manual planning based on experience data. This is not only inefficient in task planning, but may also lead to incomplete planning information due to the bias of experience data. Summary of the Invention

[0003] This application provides a task planning method, system, electronic device, storage medium, and program product to alleviate or solve one or more technical problems existing in the prior art.

[0004] In a first aspect, embodiments of this application provide a task planning method, the method comprising: receiving a planning request for an object creation task, the planning request including task description information of the object creation task; invoking an intelligent planning module to determine target creation parameters of the object creation task based on the task description information and auxiliary information associated with the object creation task, the auxiliary information being used to characterize the constraints of the object creation task; invoking an intelligent filtering module to filter at least one first resource in a resource pool that matches the target creation parameters, the first resource being used to create the object to be created corresponding to the object creation task; and generating task planning information for the object creation task based on the target creation parameters and the at least one first resource.

[0005] Secondly, embodiments of this application provide a task planning system, the system comprising: a scheduling module, an intelligent planning module, an intelligent filtering module, and an intelligent generation module; the scheduling module being configured to schedule each intelligent module upon receiving a planning request for an object creation task, the planning request including task description information of the object creation task; the intelligent planning module being configured to determine target creation parameters for the object creation task based on the task description information and auxiliary information associated with the object creation task, the auxiliary information being used to characterize the constraints of the object creation task; the intelligent filtering module being configured to filter at least one first resource in a resource pool that matches the target creation parameters, the first resource being used to create the object to be created corresponding to the object creation task; and the generation module being configured to generate task planning information for the object creation task based on the target creation parameters and the at least one first resource.

[0006] Thirdly, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored in the memory, wherein the processor implements any of the methods of embodiments of this application when executing the computer program.

[0007] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method of any one of the embodiments of this application.

[0008] Fifthly, embodiments of this application provide a computer program product, including a computer program, which, when executed by a processor, implements any of the methods described in the embodiments of this application.

[0009] In the technical solution provided in this application embodiment, upon receiving a planning request for an object creation task, an intelligent planning module is invoked to determine the target creation parameters of the object creation task, and an intelligent filtering module is invoked to filter a first resource in the resource pool that matches the target creation parameters. Based on the target creation parameters and the first resource, task planning information for the object creation task is generated. Thus, multiple intelligent modules collaborate without human intervention, not only achieving automated generation of task planning information and improving task planning efficiency, but also, since auxiliary information can characterize the constraints of the object creation task, determining the target creation parameters based on both task description information and auxiliary information improves the accuracy of the target creation parameters, thereby enhancing the accuracy of task planning.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application, it can be implemented according to the contents of the specification. In order to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below. Attached Figure Description

[0011] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments according to this application and should not be construed as limiting the scope of this application.

[0012] Figure 1 This illustration shows an application scenario diagram of the task planning method provided in an embodiment of this application; Figure 2 A flowchart of the task planning method provided in an embodiment of this application is shown; Figure 3 A schematic diagram of the task planning method provided in an embodiment of this application is shown; Figure 4This illustration shows a schematic diagram of constructing training samples according to an embodiment of this application; Figure 5 A schematic diagram of the reinforcement training process provided in an embodiment of this application is shown; Figure 6 A schematic diagram of the task planning apparatus provided in an embodiment of this application is shown; Figure 7 A block diagram of an electronic device provided in an embodiment of this application is shown. Detailed Implementation

[0013] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the concept or scope of this application. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0014] First, the terms used in the embodiments of this application are explained as follows: Supervised Fine-Tuning (SFT) trains the network to be trained using labeled data.

[0015] Reinforcement Training (RT) is a machine learning training paradigm where the model learns decision-making policies by receiving reward signals, with the goal of maximizing long-term cumulative rewards. Its core principle lies in optimizing the policy through a trial-and-error feedback mechanism, rather than relying on fixed labeled data.

[0016] Group Relative Policy Optimization (GRPO) is a reinforcement learning optimization algorithm for models. Its core mechanism is to group multiple outputs generated by the model under the same input into a group, normalize the data by calculating the relative advantage within the group, and optimize the model based on this.

[0017] To facilitate understanding of the technical solutions of the embodiments of this application, the relevant technologies of the embodiments of this application are described below. The following relevant technologies are optional solutions and can be combined with the technical solutions of the embodiments of this application in any way, and all of them fall within the protection scope of the embodiments of this application.

[0018] As mentioned earlier, the creation task is usually planned before creating digital objects. However, current task planning mainly relies on manual analysis of creation requirements and manual planning based on experience data. This is not only inefficient but also prone to incomplete planning information due to biases in the experience data. To address this technical problem, some data analysis-based tools have been proposed, but their functions focus on the presentation and statistics of historical data, still requiring users to complete the core planning and selection decisions themselves. Some matching engines have also been proposed, which perform intent recognition or simple keyword matching based on templates, but they struggle to deeply understand the complex constraints and implicit goals behind the task, leading to biases in the interpretation of the user's true intent. Some content generation tools have also been proposed, which mainly utilize large language models for text creation but do not achieve closed-loop coupling with structured resource libraries and real-time environmental conditions. Therefore, while these tools offer automation, the degree of automation is limited, and they cannot automatically complete the entire closed-loop process from intent analysis and solution planning to accurate resource matching based on the user's initial request.

[0019] In view of this, embodiments of this application provide a task planning method, system, electronic device, storage medium, and program product, which aim to improve the efficiency and accuracy of task planning, and will be described in detail below.

[0020] Figure 1 This illustration shows an application scenario diagram of a task planning method provided in an embodiment of this application, such as... Figure 1 As shown, this scenario includes both the client and the server.

[0021] The client can be a mobile phone, tablet computer, desktop computer, laptop, vehicle terminal, smart wearable device, or other terminal device. Figure 1 The example provided uses a mobile phone as the client. The client can be configured with a target application, which can be a standalone application (App), a mini-program embedded in other applications, or a webpage, etc. This target application can include a session interface where the user can interact with the client and input task description information for the object creation task. Correspondingly, the client can send a task planning request to the server based on the user-input task description information, and upon receiving the task planning information from the server, display the task planning information in the session interface. This task planning information can include target creation parameters and at least one first resource, which is used to create the object to be created corresponding to the object creation task.

[0022] The server can be a standalone physical server, a server cluster, or a cloud server for cloud computing, etc. Figure 1The example shown uses a standalone physical server as an illustration. The server can be configured with a task planning system, which may include multiple intelligent modules that collaborate to create tasks from planned objects. In the case of a server cluster, these intelligent modules can be deployed on the same server or on different servers.

[0023] In some implementations, such as Figure 1 As shown, the task planning system may include a scheduling module, an intelligent planning module, an intelligent filtering module, and an intelligent generation module. The scheduling module, upon receiving a planning request for an object creation task, schedules the various intelligent modules. The planning request includes the task description information of the object creation task. The intelligent planning module determines the target creation parameters of the object creation task based on the task description information and auxiliary information associated with the object creation task. The auxiliary information characterizes the constraints of the object creation task. The intelligent filtering module filters at least one first resource from the resource pool that matches the target creation parameters. The generation module generates task planning information for the object creation task based on the target creation parameters and at least one first resource.

[0024] In some implementations, any intelligent module can be implemented as a functional module with natural language understanding and task planning capabilities, such as an AI agent. An intelligent module is a dynamic integration of model functionality and task planning capabilities, capable of understanding received natural language instructions and performing corresponding operations based on the results of that understanding.

[0025] In some implementations, the task planning system can also be configured with multiple prompts, each used to guide the intelligent module to perform a corresponding operation. As an example, the first prompt can guide the intelligent planning module to determine the target creation parameters of the object creation task based on the task description information and auxiliary information associated with the object creation task. As another example, the second prompt can guide the intelligent filtering module to filter at least one first resource in the resource pool that matches the target creation parameters.

[0026] It should be pointed out that, Figure 1 This is for illustrative purposes only and not intended to limit the scope of the application. If the client has the capability to deploy a task planning system, task planning can also be performed by the task planning system on the client side without interaction with the server. If the server includes an editing module, it can also receive planning requests from the user without interaction with the client. The following explanation will use client-server interaction as an example.

[0027] Thus, multiple intelligent modules work together without human intervention, which not only automates the generation of task planning information and improves the efficiency of task planning, but also improves the accuracy of task planning by determining the target creation parameters based on the task description information and auxiliary information, since the auxiliary information can characterize the constraints of the object creation task.

[0028] It should be noted that the application scenarios or examples provided in the embodiments of this application are for ease of understanding, and the embodiments of this application do not specifically limit the application of the technical solutions. In addition, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0029] The technical solution of this application and how it solves the aforementioned technical problems are described in detail below with specific embodiments. The listed specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0030] Figure 2 A flowchart of a task planning method according to an embodiment of this application is shown. Figure 2 The method shown can be used by Figure 1 The task planning system shown is executed. For example... Figure 2 As shown, the method may include steps S201, S202, S203 and S204.

[0031] Step S201: Receive the planning request for the object creation task. The planning request includes the task description information of the object creation task.

[0032] In some implementations, a user can input a task description for an object creation task via a client. Based on this description, the client sends a planning request for the object creation task to the server. The server receives the planning request from the client. The object creation task can be used to create any digital object, such as a knowledge base, online store, application interface, or smart home configuration solution. The user can be the creator of the object to be created, for example, a user wanting to create an online store. The user can also initiate the planning request on behalf of the creator; for example, a user might be a member of a company (the creator) who authorizes the user to initiate the planning request to subsequently create the company's knowledge base, etc. Accordingly, the planning request may also include the creator's identification information. The task description information represents the creation intent and can be in text or voice format. If the task description information is in voice format, the server can first convert it to text format before further processing.

[0033] As one example, the task description is "I want to open an online store for original design home furnishings," meaning the object to be created for this object creation task is the online store. As another example, the task description is "Build a developer technical documentation knowledge base for project A," meaning the object to be created for this object creation task is the knowledge base.

[0034] Furthermore, as mentioned earlier, the task planning system in the server includes a scheduling module, and correspondingly, such as Figure 3 As shown, the scheduling module can receive planning requests sent by the client.

[0035] Step S202: Invoke the intelligent planning module to determine the target creation parameters of the object creation task based on the task description information and the auxiliary information associated with the object creation task. The auxiliary information is used to characterize the constraints of the object creation task.

[0036] In some implementations, after receiving a request from a client, the scheduling module in the task planning system can first parse the planning request to determine the request type. If the request type is a planning request for an object creation task, it retrieves the auxiliary information associated with the object creation task from a preset database. Then, it invokes the intelligent planning module based on the task description information and the auxiliary information. Through the intelligent planning module, the target creation parameters for the object creation task are determined based on the task description information and the auxiliary information.

[0037] In other implementations, after receiving a request from the client, the scheduling module in the task planning system can first parse the planning request to determine the request type. If the request type is a planning request for an object creation task, the intelligent planning module is invoked based on the task description information. The intelligent planning module retrieves auxiliary information associated with the object creation task from a preset database and determines the target creation parameters for the object creation task based on the task description information and the auxiliary information.

[0038] In some other implementations, the task planning system may also include a parsing module ( Figure 1 (Not shown in the image). After receiving a request from the client, the scheduling module in the task planning system can forward the request to the parsing module. The parsing module determines the request type, and if the request type is a planning request for an object creation task, it retrieves the auxiliary information associated with the object creation task from a preset database, and calls the intelligent planning module based on the task description information and the auxiliary information. The intelligent planning module then determines the target creation parameters for the object creation task based on the task description information and the auxiliary information.

[0039] Target creation parameters can be structured data (e.g., JSON format) used to characterize the features and construction requirements of the object to be created. These parameters can vary depending on the object. For example, if the object is an online store, the target creation parameters could include product category, target user group, price range, etc. As another example, if the object is a knowledge base, the target creation parameters could include knowledge type, target user group, covered chapters, knowledge depth, etc.

[0040] Step S203: Invoke the intelligent filtering module to filter at least one first resource in the resource pool that matches the target creation parameters. The first resource is used to create the object to be created corresponding to the object creation task.

[0041] After the intelligent planning module determines the target creation parameters for the object creation task, the scheduling module can invoke the intelligent filtering module to filter at least one primary resource from the resource pool that matches the target creation parameters. The resource pool can include various available creation materials, products, knowledge entries, and other resources.

[0042] The primary resource can vary depending on the object being created. As an example, if the object is an online store, the primary resource could be products that can be sold in the store, visual assets (such as store icons and product images), product descriptions, etc. As another example, if the object is a knowledge base, the primary resource could be knowledge used to populate the knowledge base, journals, books, tutorials, product documentation, code, etc.

[0043] Step S204: Generate task planning information for the object creation task based on the target creation parameters and at least one first resource.

[0044] After the intelligent filtering module obtains the first resource, the scheduling module can call the intelligent generation module. The intelligent generation module generates task planning information for the object creation task based on the target creation parameters and at least one first resource. The intelligent generation module can generate task planning information in a preset format independently based on the target creation parameters and at least one first resource. Alternatively, it can call a pre-trained generation model to generate task planning information based on the target creation parameters and at least one first resource. This task planning information integrates parameter configuration and resource selection, serving as a guiding scheme for the object creation task.

[0045] In the technical solution provided in this application embodiment, upon receiving a planning request for an object creation task, an intelligent planning module is invoked to determine the target creation parameters of the object creation task, and an intelligent filtering module is invoked to filter a first resource in the resource pool that matches the target creation parameters. Based on the target creation parameters and the first resource, task planning information for the object creation task is generated. Thus, multiple intelligent modules collaborate without human intervention, not only achieving automated generation of task planning information and improving task planning efficiency, but also, since auxiliary information can characterize the constraints of the object creation task, determining the target creation parameters based on both task description information and auxiliary information improves the accuracy of the target creation parameters, thereby enhancing the accuracy of task planning.

[0046] Considering that user-inputted task descriptions are often colloquial and incomplete, in some implementations, determining the target creation parameters of an object creation task based on the task description information and the auxiliary information associated with the object creation task may include: correcting the task description information based on the auxiliary information to obtain intermediate information, which includes the task category and initial creation parameters; and after verifying the initial creation parameters according to the verification rules corresponding to the task category, determining the verified initial creation parameters as the target creation parameters.

[0047] Since the task description information input by the user is often a brief description, in order to improve the accuracy of task planning, the intelligent planning module can correct the task description information based on auxiliary information to obtain intermediate information. Furthermore, in the correspondence between task categories and verification rules, the module obtains the verification rules corresponding to the task categories in the intermediate information. Based on the obtained verification rules, the initial creation parameters contained in the intermediate information are verified, and after successful verification, the corresponding initial creation parameters are determined as the target creation parameters. The correction can be to improve or supplement the information in the task description. The task category is a classification identifier used to distinguish different objects to be created, such as store creation and knowledge base creation. Verification rules can be set as needed in actual applications, and this application does not impose specific limitations on them.

[0048] As an example, the object to be created is an online store, and the task description includes "original design furniture." Based on the auxiliary information, "original design furniture" can be updated to "original design furniture with aesthetic style." The validation rules can be used to verify product categories, price ranges, store creation qualifications, etc.

[0049] Therefore, by correcting the task description information and verifying the initial creation parameters based on the verification rules corresponding to the task category, the compliance and rationality of the target creation parameters can be improved, thereby reducing the risk of creation failure when creating objects to be created in the future.

[0050] Considering that related technologies often directly match resources based on user-inputted task descriptions, ignoring the impact of the creator's own actions and the development trends of the relevant field, some implementation methods, such as... Figure 3 As shown, the final target creation parameters can be determined based on the creator's profile data and the domain status information of the domain to which the object to be created belongs. In other words, the aforementioned auxiliary information includes the creator's profile data and the domain status information of the domain to which the object to be created belongs. Correspondingly, the aforementioned correction of the task description information based on the auxiliary information to obtain intermediate information can include: parsing the task description information to obtain the task category and set of requirement parameters; determining the main constraints corresponding to the object creation task based on the profile data; determining the environmental constraints corresponding to the object creation task based on the domain status information; determining the initial creation parameters based on the set of requirement parameters, main constraints, and environmental constraints; and generating intermediate information based on the task category and initial creation parameters.

[0051] The intelligent planning module can invoke a semantic parsing model to parse the task description information, obtaining the task category and user requirement parameter set for the object creation task. It also retrieves corresponding profile data from the profile database based on the creator's identifier information included in the planning request. Analyzing this profile data yields the main constraints corresponding to the object creation task. Furthermore, it retrieves domain state information from the domain database corresponding to the task category and analyzes this information to obtain environmental constraints. The requirement parameter set, main constraints, and environmental constraints are then fused to obtain initial creation parameters. Finally, intermediate information is generated according to a preset format (e.g., JSON format) based on the task category and initial creation parameters.

[0052] The requirements parameter set consists of various basic requirements extracted from the task description information. Taking an online store as an example, the requirements parameter set may include user-configured product categories and quantities. Profile data comprises a series of attributes and behavioral records related to the creator, used to characterize the creator's capabilities, preferences, and historical performance. Domain state information is a collection of dynamic and static data related to the industry, market, or creation platform (the platform used to create the object), reflecting the external environment of the creation task. Subject constraints, derived from the profile data, are restrictive descriptions of the creator's capabilities, permissions, or preferences when performing the current object creation task, reflecting the boundaries of "what can be done" and "how it is inclined to be done." Environmental constraints, derived from the domain state information analysis, are external conditions that restrict or provide opportunities for task execution, reflecting the boundaries of "in what environment" and "what can or cannot be done."

[0053] Therefore, by clearly distinguishing between subject constraints and environmental constraints, the calibration process becomes more systematic and interpretable. This not only improves the rationality and feasibility of the final generated target creation parameters, but also adapts to the personalized and dynamic needs of different users in different market environments.

[0054] In some implementations, the profile data includes the creator's identity information and historical behavior data related to the object creation task. Accordingly, determining the subject constraints corresponding to the object creation task based on the profile data may include: determining the creator's creation permissions based on the identity information; determining the creator's preference information based on the historical behavior data; and determining the creation permissions and preference information as subject constraints.

[0055] The intelligent planning module can obtain the creation permissions corresponding to the creator's identity information from the mapping relationship between creation permissions and identity information. It also calls the preference analysis model to analyze the creator's historical behavioral data to obtain the creator's preference information. The creation permissions and preference information are then used as the main constraints. Identity information is used to identify and distinguish the creator, and may include their own attributes and data such as qualifications, roles, and levels granted by the creation platform. Examples include user identifiers, enterprise certification status, membership levels, and credit scores. Creation permissions are rules derived from identity information describing whether the creator has the right to perform specific operations, or within what scope they can perform those operations. Historical behavioral data includes the creator's historical operation records on the creation platform, such as historical objects created and records of resource operations (e.g., browsing, saving, purchasing). Preference information includes the creator's habits, interests, and other related information.

[0056] As an example, the planning request includes the task description: "I want to open an online store for original design home furnishings." Based on the user's membership level and credit score, the permission information retrieved indicates that the user has the authority to create furniture stores, with a maximum of five online stores. The user's historical behavior data shows that they have frequently browsed Nordic minimalist home furnishings in the past week, confirming the user's preference as Nordic minimalism.

[0057] As another example, the planning request includes a task description of writing a developer document for the product's intelligent service interface. Based on the senior engineer identity information, the user's permissions are determined to be access to key code repositories with unlimited object creation. Historical behavioral data indicates that the user's past technical documents heavily utilize charts, thus the preference information is determined to be the use of charts.

[0058] By introducing permission checks at the initial planning stage, invalid solutions that exceed the creator's capabilities can be avoided. Incorporating preference information based on historical behavioral data makes the generated task planning information more aligned with the creator's habits and preferences, thus improving the acceptability and usability of the task planning information.

[0059] In some implementations, the domain state information includes development trend information and resource evaluation information. Accordingly, the aforementioned determination of the environmental constraints corresponding to the object creation task based on the domain state information may include: determining the user group the domain is targeting and the user group's preferred resources based on the development trend information; determining the risk resources corresponding to the object creation task based on the resource evaluation information; and determining the user group, preferred resources, and risk resources as environmental constraints.

[0060] The intelligent planning module can call a domain analysis model to analyze development trend information, obtain the mainstream user group that accessed historical objects (of the same category as the object to be created) in the domain to which the object to be created belongs, and the resources that this mainstream user group generally prefers. This mainstream user group is then identified as the target user group for the domain to which the object to be created belongs, and the resources favored by this user group are determined as the user group's preferred resources. Furthermore, it analyzes resource evaluation information, identifying resources with potential problems or unsuitable for the current object creation task as risk resources.

[0061] Among these, development trend information is used to characterize the direction of dynamic changes in the field, such as search hot keywords, resource categories with rapidly growing purchase rates, social media discussion popularity, technology stack adoption trends, accumulated reusable experience data, and seasonal demand changes such as holidays or seasonal changes. Resource evaluation information can include the search popularity of resource categories, return rates of resources (such as products), whether resources (such as products) have infringement risks, whether resources (such as products) have slow sales risks, the completion rate of reading resources (such as knowledge documents), the number of resource collections and forwards, and whether resources have hidden security risks. Preference resources are resource types or specific resource instances that are highly relevant to the user group identified by the analysis and currently show a positive development trend, and can be regarded as opportunities or popular options. Risky resources are resource types or specific resource instances identified based on resource evaluation information that have potential problems or are not suitable for recommendation in the current task, and can be regarded as items to be avoided.

[0062] As an example, the planning request included the task description: "I want to open an online store for original design furniture." Trend information indicates that search terms such as "designer furniture" and "niche home furnishings" have seen a growth rate exceeding 80% in the past 30 days. Among social media discussions in this field, the core users are urban white-collar workers and freelancers who value aesthetics in their lifestyles. Products tagged with "sustainable materials" (such as recycled plastics and certified wood) and "modular design" have significantly higher conversion rates than average. Resource evaluation information indicates that some original design products are marked as high-risk due to high rates of damage during shipping caused by insufficient packaging or returns due to unclear size labeling. Based on this, it can be determined that the target user group for this field is urban residents with high aesthetic demands; preferred resources are sustainable material furniture and modular furniture; risky resources are products with simple packaging and incomplete size information.

[0063] Therefore, by identifying the target user group and its preferences, the generated task planning information is more likely to meet market demands and increase the success rate; by identifying risky resources, potential problems can be eliminated in advance during the planning stage, thereby improving the robustness and quality of task planning information.

[0064] Furthermore, the aforementioned determination of initial creation parameters based on the set of requirement parameters, main constraints, and environmental constraints may include: for any requirement parameter in the set of requirement parameters, updating the requirement parameter if the requirement parameter does not match the main constraints and / or environmental constraints; and determining the updated set of requirement parameters and the supplementary parameters as the initial creation parameters if the main constraints and / or environmental constraints contain supplementary parameters and the updated set of requirement parameters does not contain supplementary parameters.

[0065] The intelligent planning module compares each requirement parameter in the requirement parameter set with the main constraints and environmental constraints. If a parameter does not match any constraint (e.g., the requirement specifies an option marked as a risky resource by the environmental constraints), the requirement parameter can be updated based on the constraints (e.g., replacing the risky resource with a preferred resource) or based on preset rules (e.g., adjusting a value exceeding the permission range to the maximum allowed value). After verifying all requirement parameters in the requirement parameter set and updating any necessary parameters, an updated requirement parameter set is obtained. The intelligent planning module checks whether the main constraints and environmental constraints contain supplementary parameters that should be used as creation parameters. If supplementary parameters are included, they are combined with the updated requirement parameter set to determine the initial creation parameters. For example, if the supplementary parameters for "required qualifications" in the main constraints or "risky resources" in the environmental constraints are not yet included in the updated requirement parameter set, they are combined with the updated requirement parameter set to determine the initial creation parameters.

[0066] For example, if the requirement parameter "industrial style" does not match "Nordic minimalism" in the main constraints, an update is triggered. The intelligent planning module can update "industrial style" to a compromise or a more user-preferred "Nordic industrial eclectic style" based on rules or models. If the requirement parameter "solid wood" partially matches "risky materials" in the environmental constraints, the system will update it to "certified sustainable solid wood" to mitigate risks.

[0067] Therefore, by implementing mismatch detection and automatic updates, the need for manual intervention is reduced, and the feasibility and automation of task planning information are improved. The introduction of supplementary parameters ensures that the initial creation parameters contain all the key information required to perform the object creation task, which helps improve the accuracy of subsequent resource selection.

[0068] To improve resource screening efficiency, such as Figure 3As shown, the intelligent filtering module can extract first filtering information through the extraction model and filter first resources based on the first filtering information. That is, the aforementioned filtering of at least one first resource in the resource pool that matches the target creation parameters can include: calling the extraction model to determine the first filtering information corresponding to the target creation parameters; and filtering at least one first resource in the resource pool that matches the first filtering information.

[0069] In some implementations, the intelligent filtering module can input target creation parameters into a pre-trained extraction model, analyze the target creation parameters through the extraction model, and determine first filtering information. This first filtering information can come from the target creation parameters or other information determined based on the target creation parameters. Based on the first filtering information, a filtering interface is invoked to filter first resources from the resource pool that match the first filtering information, or a filtering request is sent to the resource pool management module based on the first filtering information, and the module receives first resources that match the first filtering information from the management module.

[0070] For example, the target creation parameters include {"Store Positioning": "Original Design Home Furnishings", "Style": "Nordic-Japanese Mix", "Core Materials": "Wood, Cotton & Linen", "Price Range": "Mid-to-High-End", "Target Scene": "Small Apartment Living Room" and "Bedroom Storage"}. The first filter information determined based on the target creation parameters includes wood design home furnishings, Japanese-Nordic style, small apartment solid wood furniture, cotton and linen home decor, designer brands, price range of 500-5000, and tags such as "original" and / or "designer". Designer-designed wood furniture and cotton and linen products that match this first filter information and are selected from the resource pool are used as the first resource.

[0071] Since the target creation parameters are oriented towards planning and definition, while the retrieval of the resource pool usually has its own query language or indexing method, extracting the model to determine the first filtering information corresponding to the target creation parameters, and filtering the first resource based on the first filtering information, can better fit the retrieval characteristics of the resource pool and improve the relevance of the recalled resources.

[0072] To improve the training efficiency of the extraction module, in some implementations, the method further includes: for any historical object among multiple historical objects, calling the teacher model to determine the second screening information corresponding to the historical creation parameters of the historical object; using the training sample set corresponding to multiple historical objects to train the student model to obtain the extraction model, wherein any training sample in the training sample set includes the historical creation parameters and the corresponding second screening information.

[0073] In some implementations, such as Figure 4As shown, for each of multiple historical objects that have been successfully created and verified (e.g., popular online stores, high-quality document knowledge bases, etc.), the historical creation parameters of that historical object can be obtained. Furthermore, based on RAG (Retrieval Augmented Generation), knowledge such as domain experience and creation rules corresponding to the historical object can be retrieved. These historical creation parameters and the retrieved knowledge are input into a teacher model to generate second filtering information corresponding to the historical object. At least one second resource matching the second filtering information is selected from the resource pool, and a relevance score between the second resource and the historical object is determined. If the relevance score is greater than a threshold, training samples are constructed based on the historical creation parameters and the second filtering information of the historical object; that is, the second filtering information serves as the label for the historical creation parameters. Each constructed training sample is used as a training sample set to train the student model, resulting in an extraction model. The process of determining the relevance score can involve inputting the description information of the historical object and the description information of the second resource into a first scoring model, which then outputs the relevance score between the second resource and the historical object.

[0074] Furthermore, if the relevance score is not greater than the score threshold, test samples can be constructed based on the historical creation parameters and second screening information of the historical object to obtain a test sample library, and the extraction model can be tested using the test sample library.

[0075] The teacher model can be a large or complex language model with strong generalization and reasoning capabilities, used to generate high-quality secondary filtering information relevant to the actual success of historical objects. However, it is often computationally expensive and slow to respond, making it unsuitable for direct deployment in online services. Therefore, this powerful teacher model is used to generate high-quality secondary filtering information for historical objects as annotations. Then, using these labeled data (i.e., training samples), a lighter and more efficient "student model" is trained. This student model learns to mimic the teacher model's ability to extract filtering information from creation parameters, resulting in an extraction model. In other words, the student model is a relatively simple model to be trained, and the training goal is to make its output close to that of the teacher model, while possessing faster reasoning speed and lower resource consumption.

[0076] Therefore, the reasoning ability of the teacher model is used to automatically generate training labels, avoiding a large amount of manual annotation work. By distilling the capabilities of the teacher model into a lighter student model, the final extraction model can meet the efficiency requirements of online services while maintaining high performance.

[0077] To enable cross-module attribution, reverse optimization of the information selection strategy is needed to reduce error propagation where errors in one module lead to errors in another, and a trade-off is made under a global objective. In some implementations, supervised training and reinforcement training are combined to train the extraction model. That is, the aforementioned method of training the student model using training sample sets corresponding to multiple historical objects to obtain the extraction model can be further divided into: supervising the student model using each training sample in the training sample set to obtain an intermediate model; and using the historical creation parameters in the training sample set to perform reinforcement training on the intermediate model to obtain the extraction model.

[0078] A standard supervised learning algorithm (e.g., gradient descent) can be used to supervise the training model, taking the training samples as input and the second selection information generated by the teacher model as the expected output, until the loss function converges, resulting in an intermediate model. This intermediate model can generate selection information similar to that generated by the teacher model. Then, using the historical creation parameters from the training sample set, the intermediate model is reinforced to obtain the extraction model.

[0079] In some implementations, the intermediate model is reinforced and trained using historical creation parameters from the training sample set to obtain the extraction model. This may include: generating multiple third-level filtering information for any historical creation parameter in the training sample set using the intermediate model; obtaining at least one third-level resource from the resource pool that matches any third-level filtering information for any given third-level filtering information; determining a comprehensive matching score for the third-level filtering information based on the third-level filtering information and the at least one third-level resource, where the comprehensive matching score characterizes the degree of comprehensive matching between the third-level filtering information and the at least one third-level resource and the corresponding historical object; and adjusting the model parameters of the intermediate model based on the comprehensive matching scores of the multiple third-level filtering information until the training stopping condition is met to obtain the extraction model.

[0080] In intensive training, such as Figure 5As shown, for each training sample in the training sample set, the historical creation parameters are input into the intermediate model. The intermediate model samples from the generated probability distribution to obtain multiple third-selection information and output N third-selection information, where N is an integer greater than 1. The probability distribution includes the probabilities of multiple candidate selection information. For each of the multiple third-selection information corresponding to the historical creation parameters, at least one third resource matching that third-selection information is obtained from the resource pool. A second scoring model is then called to determine the first matching score between the third-selection information and the historical object, and the second matching score between the third resource and the historical object. The first and second matching scores are weighted and summed according to preset weights to obtain the comprehensive matching score of the third-selection information. Based on the comprehensive matching scores of each third-selection information, the model parameters of the intermediate model are adjusted according to the GRPO strategy until the training stopping condition is met, resulting in the extraction model.

[0081] The first matching score characterizes the degree of matching between the third filtering information and the historical object, while the second matching score characterizes the degree of matching between the third resource and the historical object. The first matching score is determined based on scoring dimensions such as the length of the third filtering information, whether there is redundancy among the filtering fields in the third filtering information, and whether the third filtering information is relevant to the corresponding historical object. The second matching score is determined based on scoring dimensions such as whether the third resource is relevant to the corresponding historical object, the diversity of the third resource, and the rationality of the combination of the third resources. The process of adjusting model parameters based on the GRPO strategy can refer to relevant technologies, and is not specifically limited in this application.

[0082] Therefore, in reinforcement training, by combining the quality of the third-party screening information and the quality of the recalled third-party resources, a comprehensive matching score is determined and used to adjust model parameters. This provides the model with a more refined and stable learning signal, reducing training instability. By generating and scoring multiple third-party screening information for the same input, the model is encouraged to explore, helping to discover screening strategies that are not known to the teacher model but have better actual results, thereby improving the performance of the trained extraction model.

[0083] Corresponding to the application scenarios and methods provided in the embodiments of this application, the embodiments of this application also provide a task planning device, which can be applied to... Figure 1 The server shown is as follows. Figure 6 As shown, the device includes: The receiving unit 601 is used to receive a planning request for an object creation task, wherein the planning request includes task description information of the object creation task; The determining unit 602 is used to call the intelligent planning module to determine the target creation parameters of the object creation task based on the task description information and the auxiliary information associated with the object creation task. The auxiliary information is used to characterize the constraints of the object creation task. The filtering unit 603 is used to call the intelligent filtering module to filter at least one first resource in the resource pool that matches the target creation parameters. The first resource is used to create the object to be created corresponding to the object creation task. The generation unit 604 is used to generate task planning information for the object creation task based on the target creation parameters and the at least one first resource.

[0084] In some implementations, the determining unit 602 is specifically used to: correct the task description information based on the auxiliary information to obtain intermediate information, the intermediate information including task category and initial creation parameters; and after the initial creation parameters are verified according to the verification rules corresponding to the task category, determine the verified initial creation parameters as the target creation parameters.

[0085] In some implementations, the auxiliary information includes profile data of the creator of the object to be created and domain status information of the domain to which the object to be created belongs. The determining unit 602 is further specifically used for: parsing the task description information to obtain the task category and the set of requirement parameters; determining the subject constraints corresponding to the object creation task based on the profile data; determining the environmental constraints corresponding to the object creation task based on the domain status information; determining the initial creation parameters according to the set of requirement parameters, the subject constraints, and the environmental constraints; and generating the intermediate information according to the task category and the initial creation parameters.

[0086] In some implementations, the profile data includes identity information and historical behavior data related to the object creation task, and the determining unit 602 is further specifically used to: determine the creator's creation permission based on the identity information; determine the creator's preference information based on the historical behavior data; and determine the creation permission and the preference information as the subject constraint conditions.

[0087] In some implementations, the domain state information includes development trend information and resource evaluation information. The determining unit 602 is further specifically used to: determine the user group targeted by the domain and the preferred resources of the user group based on the development trend information; determine the risk resources corresponding to the object creation task based on the resource evaluation information; and determine the user group, the preferred resources, and the risk resources as the environmental constraints.

[0088] In some embodiments, the determining unit 602 is further specifically configured to: update the requirement parameter if the requirement parameter does not match the main constraints and / or the environmental constraints for any requirement parameter in the requirement parameter set; and determine the updated requirement parameter set and the supplementary parameter as the initial creation parameters if the main constraints and / or the environmental constraints contain supplementary parameters and the updated requirement parameter set does not contain the supplementary parameters.

[0089] In some implementations, the filtering unit 603 is specifically used to: invoke the extraction model to determine the first filtering information corresponding to the target creation parameters; and filter at least one first resource in the resource pool that matches the first filtering information.

[0090] In some embodiments, the device further includes a training module for calling a teacher model for any one of a plurality of historical objects to determine the second filtering information corresponding to the historical creation parameters of the historical object; and training a student model using a training sample set corresponding to the plurality of historical objects to obtain the extraction model, wherein any training sample in the training sample set includes the historical creation parameters and the corresponding second filtering information.

[0091] In some implementations, the training module is specifically used to: perform supervised training on the student model using each training sample in the training sample set to obtain an intermediate model; and perform reinforcement training on the intermediate model using each historical creation parameter in the training sample set to obtain the extraction model.

[0092] In some implementations, the training module is further specifically used for: generating multiple third-level filtering information for any historical creation parameter in the training sample set through the intermediate model; for any third-level filtering information, obtaining at least one third-level resource from the resource pool that matches the third-level filtering information; determining a comprehensive matching score for the third-level filtering information based on the third-level filtering information and the at least one third-level resource, wherein the comprehensive matching score is used to characterize the comprehensive matching degree between the third-level filtering information and the at least one third-level resource and the corresponding historical object; and adjusting the model parameters of the intermediate model based on the comprehensive matching scores of the multiple third-level filtering information until the training stopping condition is met, thereby obtaining the extraction model.

[0093] The functions of each module in the devices of this application embodiment can be found in the corresponding descriptions of the methods described above, and they have corresponding beneficial effects, which will not be repeated here. Furthermore, the device embodiments described above are merely illustrative. The modules described as separate components may or may not be physically separate. The components illustrated as modules may or may not be physical modules; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this application solution according to actual needs.

[0094] Figure 7 This is a block diagram of an electronic device used to implement embodiments of this application. For example... Figure 7 As shown, the electronic device includes a memory 701 and a processor 702. The memory 701 stores a computer program that can run on the processor 702. When the processor 702 executes the computer program, it implements the method described in the above embodiments. The number of memories 701 and processors 702 can be one or more. In a specific implementation, the electronic device may also include a communication interface 703 for communicating with external devices and performing data exchange and transmission.

[0095] In practical implementation, if the memory 701, processor 702, and communication interface 703 are implemented independently, they can be interconnected via a bus to communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0096] Optionally, in a specific implementation, if the memory 701, processor 702, and communication interface 703 are integrated on a single chip, the memory 701, processor 702, and communication interface 703 can communicate with each other through an internal interface.

[0097] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method provided in this application.

[0098] This application provides a computer program product, including a computer program that, when executed by a processor, implements the method provided in this application.

[0099] This application also provides a chip including a processor for calling and executing instructions stored in a memory, causing a communication device with the chip installed to perform the method provided in this application.

[0100] This application also provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, output interface, processor, and memory are connected through an internal connection path. The processor is used to execute code in the memory. When the code is executed, the processor is used to execute the method provided in the application embodiment.

[0101] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting Advanced Reduced Instruction Set Machines (ARM) architecture.

[0102] Further, optionally, the aforementioned memory may include read-only memory and random access memory. The memory may be volatile memory or non-volatile memory, or may include both. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Sync Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0103] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0104] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0105] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0106] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0107] The logic and / or steps described in the flowchart or otherwise herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0108] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, the program being stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiments.

[0109] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. This storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0110] The above description is merely an exemplary embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope described in this application, and these should all be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A task planning method, the method comprising: Receive a planning request for an object creation task, the planning request including task description information of the object creation task; The intelligent planning module is invoked to determine the target creation parameters of the object creation task based on the task description information and the auxiliary information associated with the object creation task. The auxiliary information is used to characterize the constraints of the object creation task. The intelligent filtering module is invoked to filter at least one first resource in the resource pool that matches the target creation parameters. The first resource is used to create the object to be created corresponding to the object creation task. Based on the target creation parameters and the at least one first resource, generate task planning information for the object creation task.

2. The method according to claim 1, wherein, The step of determining the target creation parameters of the object creation task based on the task description information and the auxiliary information associated with the object creation task includes: Based on the auxiliary information, the task description information is corrected to obtain intermediate information, which includes the task category and initial creation parameters. After the initial creation parameters pass the verification according to the verification rules corresponding to the task category, the verified initial creation parameters are determined as the target creation parameters.

3. The method according to claim 2, wherein, The auxiliary information includes the profile data of the creator of the object to be created and the domain status information of the domain to which the object to be created belongs. The step of correcting the task description information based on the auxiliary information to obtain intermediate information includes: Parse the task description information to obtain the task category and set of requirement parameters; Based on the profile data, determine the main constraints corresponding to the object creation task; Based on the domain state information, determine the environmental constraints corresponding to the object creation task; Based on the set of requirement parameters, the main constraints, and the environmental constraints, determine the initial creation parameters; The intermediate information is generated based on the task category and the initial creation parameters.

4. The method according to claim 3, wherein, The profile data includes identity information and historical behavior data related to the object creation task. The step of determining the subject constraints corresponding to the object creation task based on the profile data includes: Based on the identity information, determine the creator's creation permissions; Based on the historical behavior data, determine the creator's preference information; The creation permission and the preference information are determined as the subject constraints.

5. The method according to claim 3, wherein, The domain status information includes development trend information and resource evaluation information. The step of determining the environmental constraints corresponding to the object creation task based on the domain status information includes: Based on the development trend information, determine the target user group for the field and the preferred resources of the user group; Based on the resource evaluation information, determine the risk resources corresponding to the object creation task; The user group, the preference resources, and the risk resources are defined as the environmental constraints.

6. The method according to claim 3, wherein, The step of determining the initial creation parameters based on the set of requirement parameters, the main constraints, and the environmental constraints includes: For any requirement parameter in the set of requirement parameters, if the requirement parameter does not match the main constraints and / or the environmental constraints, the requirement parameter is updated. If the main constraints and / or the environmental constraints contain supplementary parameters, and the updated set of requirements parameters does not contain the supplementary parameters, the updated set of requirements parameters and the supplementary parameters shall be determined as the initial creation parameters.

7. The method according to any one of claims 1 to 6, wherein, The step of filtering at least one first resource in the resource pool that matches the target creation parameters includes: Invoke the extraction model to determine the first filtering information corresponding to the target creation parameters; In the resource pool, at least one first resource that matches the first filtering information is selected.

8. The method according to claim 7, wherein, The method further includes: For any one of the multiple historical objects, call the teacher model to determine the second filtering information corresponding to the historical creation parameters of the historical object; The student model is trained using the training sample set corresponding to the multiple historical objects to obtain the extraction model. Each training sample in the training sample set includes historical creation parameters and corresponding second filtering information.

9. The method according to claim 8, wherein, The step of training the student model using the training sample set corresponding to the multiple historical objects to obtain the extraction model includes: The student model is trained under supervision using each training sample in the training sample set to obtain an intermediate model; The intermediate model is reinforced using the historical creation parameters in the training sample set to obtain the extracted model.

10. The method according to claim 9, wherein, The step of using historical creation parameters from the training sample set to enhance the intermediate model and obtain the extraction model includes: For any historical creation parameter in the training sample set, the intermediate model generates multiple third-stage filtering information for the historical creation parameter. For any third filter information, at least one third resource matching the third filter information is obtained from the resource pool; Based on the third screening information and the at least one third resource, a comprehensive matching score for the third screening information is determined. The comprehensive matching score is used to characterize the degree of comprehensive matching between the third screening information and the at least one third resource and the corresponding historical object. Based on the combined matching score of the multiple third-screening information, the model parameters of the intermediate model are adjusted until the training stopping condition is met, thus obtaining the extracted model.

11. A task planning system, the system comprising: The system comprises a scheduling module, an intelligent planning module, an intelligent filtering module, and an intelligent generation module. The scheduling module is used to schedule each intelligent module upon receiving a planning request for an object creation task, wherein the planning request includes task description information of the object creation task. The intelligent planning module is used to determine the target creation parameters of the object creation task based on the task description information and the auxiliary information associated with the object creation task. The auxiliary information is used to characterize the constraints of the object creation task. The intelligent filtering module is used to filter at least one first resource in the resource pool that matches the target creation parameters, and the first resource is used to create the object to be created corresponding to the object creation task. The intelligent generation module is used to generate task planning information for the object creation task based on the target creation parameters and the at least one first resource.

12. An electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein the processor, when executing the computer program, implements the method of any one of claims 1 to 10.

13. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method of any one of claims 1 to 10.

14. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 10.

Images