Demand topology graph generation method, apparatus and equipment

By using automated semantic parsing and logical relationship recognition technologies, the problem of low efficiency in demand topology graph generation has been solved, achieving efficient and reliable demand topology graph generation and improving modeling efficiency and traceability in systems engineering.

CN123068532APending Publication Date: 2026-11-10SUZHOU TONGYUAN SOFT CONTROL INFORMATION TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611292486.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-25
Publication Date
2026-11-10

AI Technical Summary

Technical Problem

The existing requirement topology diagram generation process relies on manual operation, resulting in low generation efficiency. Furthermore, different professionals have different understandings of the requirements, leading to inconsistent interpretations of the requirements.

Method used

By acquiring a unified requirement text, a pre-defined domain terminology library, and a pre-defined language model for semantic analysis, the system automatically identifies requirement entities and logical relationships, scores them in conjunction with a set of standard answers, and generates a target requirement topology map.

Benefits of technology

It achieves high-precision semantic parsing and logical relationship recognition of complex requirement texts, improves the efficiency and reliability of requirement topology graph generation, and ensures the traceability and consistency of the generated results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123068532A_ABST
    Figure CN123068532A_ABST
Patent Text Reader

Abstract

This application discloses a method, apparatus, and device for generating demand topology diagrams, relating to the field of automated topology diagram generation technology. The method includes: acquiring a unified demand text, a preset domain terminology library, and a set of standard answers; performing semantic analysis based on the unified demand text, the preset domain terminology library, and a preset language model to obtain demand entities and structured demand units; identifying logical relationships based on the demand entities and the structured demand units to obtain a set of demand logical relationships; comparing and scoring the set of demand logical relationships with the set of standard answers to obtain an evaluation score; and generating a target demand topology diagram based on the demand entities, the structured demand units, and the set of demand logical relationships when the evaluation score is greater than a preset threshold. Through semantic analysis and logical relationship identification mechanisms, the automated generation of demand topology diagrams can be achieved, improving the efficiency of demand topology diagram generation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of automated topology graph generation technology, and in particular to a method, apparatus and equipment for generating demand topology graphs. Background Technology

[0002] As the scale of complex equipment systems, industrial software systems, and intelligent manufacturing systems continues to expand, the number of demands, the hierarchy of demands, and the logical relationships between demands in the systems engineering R&D process are growing rapidly.

[0003] However, in practical engineering applications, requirements information typically exists in natural language in requirements specifications, design documents, interface documents, test specifications, and historical project documents. Existing requirements modeling methods mainly rely on manual reading of requirements documents and the manual construction of SysML (Systems Modeling Language) requirements topology diagrams or requirements relationship models by systems engineers. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, and device for generating demand topology graphs, aiming to solve the technical problem of low generation efficiency caused by the reliance on manual operation for semantic understanding and logical recognition in the current demand topology graph generation process.

[0005] To achieve the above objectives, this application proposes a method for generating a demand topology graph, the method comprising: Obtain a unified requirement text, a pre-defined domain terminology library, and a set of standard answers; Semantic parsing is performed based on the unified requirement text, the preset domain terminology library, and the preset language model to obtain requirement entities and structured requirement units; Logical relationships are identified based on the demand entities and the structured demand units to obtain a set of demand logical relationships; The set of logical relationships of the requirements and the set of standard answers are compared and scored to obtain an evaluation score; When the evaluation score is greater than a preset threshold, a target demand topology map is generated based on the demand entity, the structured demand unit, and the set of demand logical relationships.

[0006] In one embodiment, the step of identifying logical relationships based on the demand entity and the structured demand unit to obtain a set of demand logical relationships includes: Logical relationships are identified based on the demand entities and the structured demand units, and the identification results of each relationship are determined. Calculate the confidence level of the relationship identification results to obtain the confidence level corresponding to each relationship identification result; The set of logical relationships of the requirements is obtained based on the relationship identification results and the confidence levels corresponding to the relationship identification results.

[0007] In one embodiment, the step of identifying logical relationships based on the demand entity and the structured demand unit, and determining the identification results of each relationship, includes: Obtain the preset domain knowledge rule base and the text features in the unified requirement text; Based on the text features, candidate demand relationship pairs are generated based on the structured demand units; The structured demand units and the candidate demand relationship pairs are identified and processed to obtain relationship clues and semantic determination results; Intermediate results are obtained by performing relationship matching on the required entities using the domain knowledge rule base; The identification results of each relation are determined based on the candidate demand relation pairs, the relation clues, the semantic determination results, the intermediate results, and the domain knowledge rule base.

[0008] In one embodiment, the preset language model includes at least an entity recognition model and a semantic correction model. The step of performing semantic parsing based on the unified requirement text, the preset domain terminology database, and the preset language model to obtain requirement entities and structured requirement units includes: Based on the preset domain terminology library, the different entity representations in the unified requirement text are mapped to unified entity representations; The unified requirement text is input into the entity recognition model for entity recognition to obtain the requirement entity; The unified requirement text is input into the semantic correction model for semantic correction, and the correction result is obtained. A structured requirement text is obtained based on the unified entity representation, the requirement entity, and the correction result. The structured requirement text is parsed to obtain structured requirement units.

[0009] In one embodiment, the step of inputting the unified requirement text into the entity recognition model for entity recognition to obtain the requirement entity includes: Obtain the entity features of the unified requirement text; The unified requirement text is input into the entity recognition model for entity recognition, and the context entity index and entity pointer expression are obtained. Based on the entity features, the index expression, and the context entity index, a candidate entity and its corresponding confidence score are determined. The required entity is obtained based on the candidate entity and the confidence score corresponding to the candidate entity.

[0010] In one embodiment, the step of inputting the unified requirement text into the semantic correction model for semantic correction to obtain the correction result includes: The unified requirement text is input into the semantic correction model for semantic anomaly detection to obtain the anomaly detection result. When the anomaly determination result indicates that the unified requirement text has a semantic clarity problem, a first correction suggestion is output based on the text content corresponding to the semantic clarity problem. The first correction suggestion includes at least one of the following: prompting for supplementation of specific parameters, prompting for supplementation of threshold range, prompting for supplementation of applicable conditions, and prompting for supplementation of verification standards. When the anomaly determination result indicates that there is a logical consistency problem in the unified requirement text, a second correction suggestion is output based on the text content corresponding to the logical consistency problem. The second correction suggestion includes at least the conflicting requirement number, conflicting field, conflicting reason, and conflict correction suggestion. The correction result is obtained based on the first correction suggestion and / or the second correction suggestion.

[0011] In one embodiment, the step of generating a target demand topology map based on the demand entity, the structured demand unit, and the set of demand logical relationships when the evaluation score is greater than a preset threshold includes: When the evaluation score is greater than a preset threshold, a demand meta-model is established based on the demand entity, the structured demand unit, and the set of demand logical relationships. The system modeling language requirement elements are determined based on the requirement meta-model and the set of requirement logical relationships. An initial system modeling language model is generated based on the aforementioned requirement meta-model and the requirement elements of the system modeling language. The initial system modeling language model is subjected to rule detection, and the initial system modeling language model that passes the rule detection is used as the target system modeling language model; Generate a target requirement topology diagram based on the target system modeling language model.

[0012] In one embodiment, the step of determining the system modeling language requirement elements based on the requirement meta-model and the set of requirement logical relationships includes: Iterate through each demand entity in the demand meta-model and create the demand node corresponding to the demand entity; Write the field information in the demand entity to the demand node corresponding to the demand entity to obtain the first updated demand node; Traverse the relationship types in the set of requirement logical relationships, and determine the connection relationships between the first updated requirement nodes based on the relationship types; Based on the first updated demand node and the connection relationship, the second updated demand node is obtained; The second updated demand nodes are grouped to obtain the third updated demand nodes; The system modeling language requirement elements are determined based on the third updated requirement node.

[0013] Furthermore, to achieve the above objectives, this application also proposes a demand topology graph generation apparatus, which includes: The acquisition module is used to acquire standardized requirement text, a preset domain terminology library, and a set of standard answers. The processing module is used to perform semantic parsing based on the unified requirement text, the preset domain terminology library, and the preset language model to obtain requirement entities and structured requirement units. The processing module is also used to identify logical relationships based on the demand unit and the structured demand unit to obtain a set of demand logical relationships; The processing module is also used to compare and score the set of logical relationships of requirements and the set of standard answers to obtain an evaluation score; The generation module is used to generate a target demand topology map based on the demand entity, the structured demand unit, and the set of demand logical relationships when the evaluation score is greater than a preset threshold.

[0014] In addition, to achieve the above objectives, this application also proposes a demand topology graph generation device, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the demand topology graph generation method as described above.

[0015] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the demand topology graph generation method described above.

[0016] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the demand topology graph generation method described above.

[0017] One or more technical solutions proposed in this application have at least the following technical effects: 1) By using a unified requirement text, a pre-defined domain terminology library, and a pre-defined language model for automated semantic parsing, high-precision semantic parsing of complex requirement texts can be achieved.

[0018] 2) By combining automated semantic parsing and automated logical relationship recognition mechanisms, complex requirements are transformed into standardized requirement data, thereby achieving seamless integration with the generation of the target requirement topology map and improving the efficiency of target requirement topology map generation. Attached Figure Description

[0019] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a flowchart illustrating an embodiment of the method for generating a topology diagram for the requirements of this application. Figure 2 A schematic diagram comparing the conventional route and the route of the present invention provided in an embodiment of the topology map generation method required by this application; Figure 3 Example diagram of logical relationship identification provided for an embodiment of the topology graph generation method required by this application; Figure 4 Verification cases and indicator diagrams provided for an embodiment of the topology graph generation method of this application; Figure 5 This is a flowchart illustrating the second embodiment of the method for generating the topology diagram required by this application. Figure 6 A schematic diagram of the semantic parsing technology architecture provided for an embodiment of the topology graph generation method required by this application; Figure 7 This is a flowchart illustrating the third embodiment of the method for generating a topology diagram for the requirements of this application. Figure 8 A schematic diagram of consistency and conflict detection provided for an embodiment of the topology graph generation method required by this application; Figure 9 Example diagram of system output results provided for an embodiment of the topology graph generation method of this application; Figure 10 This is a timing diagram illustrating the generation of a demand topology graph, provided as an embodiment of the demand topology graph generation method of this application. Figure 11This is a schematic diagram of the module structure of the topology graph generation device according to an embodiment of this application; Figure 12 This is a schematic diagram of the device structure of the hardware operating environment involved in the requirement topology graph generation method in this application embodiment.

[0022] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0023] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0024] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0025] The main solution of this application embodiment is as follows: Obtain a unified requirement text, a preset domain terminology library, and a set of standard answers; perform semantic parsing based on the unified requirement text, the preset domain terminology library, and a preset language model to obtain requirement entities and structured requirement units; identify logical relationships based on the requirement entities and the structured requirement units to obtain a set of requirement logical relationships; compare and score the set of requirement logical relationships with the set of standard answers to obtain an evaluation score; when the evaluation score is greater than a preset threshold, generate a target requirement topology map based on the requirement entities, the structured requirement units, and the set of requirement logical relationships.

[0026] Because existing technology relies on system engineers to manually analyze requirement texts to understand the semantics of requirements, for complex requirement documents, engineers need to manually identify functional subjects, constraints, performance indicators, and logical relationships between requirements. This approach is usually inefficient, and different professionals may have different understandings of requirements, which can easily lead to inconsistent interpretations of requirements.

[0027] This application provides a solution that utilizes a pre-defined domain terminology library and a pre-defined language model to automate the semantic parsing of standardized requirement text. The parsed requirement entities and structured requirement units are combined to identify logical relationships, resulting in a set of requirement logical relationships. A pre-constructed set of standard answers is compared and scored against this set of requirement logical relationships. Only when the score meets the requirements is the target requirement topology diagram generated. This process improves the efficiency of target requirement topology diagram generation while ensuring its reliability.

[0028] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or a demand topology graph generation system capable of performing the above functions. The following description uses a demand topology graph generation system as an example to illustrate this embodiment and the subsequent embodiments. All actions involving the acquisition of signals, information, or data in this application are performed in accordance with the relevant data protection regulations of the country where the application is located and with authorization from the owner of the corresponding device.

[0029] Based on this, embodiments of this application provide a method for generating a demand topology graph, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the topology graph generation method for this application.

[0030] In this embodiment, the demand topology graph generation method includes steps S10 to S50: Step S10: Obtain a standardized requirement text, a pre-defined domain terminology library, and a set of standard answers.

[0031] It should be noted that standardized requirement text is usually obtained by preprocessing the original requirement text (usually natural language). A pre-defined domain terminology library typically refers to a structured data collection with domain terms as entries. Each entry usually includes the term name, part of speech, corresponding standard entity category, possible aliases / abbreviations, and semantic relationships with other terms (such as hyponymy, synonymy, compositional relationships). The pre-defined domain terminology library is usually the product of formalizing and digitizing domain expert knowledge and project vocabulary standards. A standard answer set is usually a manually annotated sample set. The process of obtaining a standard answer set can be as follows: sample text is extracted from requirement specifications, interface documents, test specifications, and historical requirement documents, and domain experts manually annotate the requirement entities, entity types, logical relationship types, and relationship directions to form the standard answer set.

[0032] Understandably, since the original requirement text has problems such as unstructured format, complex terminology, implicit logic, and inconsistent expression, it is necessary to first preprocess and unify the original requirement text: perform sentence segmentation, paragraphing, number recognition, terminology standardization, and stop word filtering on the requirement document to eliminate semantic differences caused by different expression methods.

[0033] like Figure 2 As shown, Figure 2This diagram illustrates the comparison between the traditional approach and the approach of this invention. The traditional approach typically involves inputting original requirement documents that rely on expert experience, followed by manual reading and understanding of the documents, manual requirement decomposition, logical relationship identification, SysML requirement model building, and consistency and conflict checks, ultimately outputting a requirement model. This manual approach results in low modeling efficiency and weak traceability. In contrast, the approach of this invention preprocesses and standardizes the input multi-source requirement documents using a domain terminology library, then performs semantic parsing and structured expression using a domain knowledge rule base. It automatically identifies requirement logical relationships, automatically generates SysML requirement models, automatically constructs requirement topology diagrams, and automatically detects consistency and conflicts, finally outputting a standardized model and a detection report. This approach offers higher modeling efficiency and greater flexibility compared to the traditional approach.

[0034] By pre-acquiring standardized requirement text, a pre-defined domain terminology library, and a set of standard answers, the subsequent semantic parsing and logical relationship recognition processes can incorporate this foundational data, laying the data foundation for the subsequent automatic generation process from requirement text to target requirement topology diagram.

[0035] Step S20: Based on the unified requirement text, the preset domain terminology library, and the preset language model, perform semantic parsing to obtain requirement entities and structured requirement units.

[0036] It should be noted that a demand entity typically refers to the basic element that constitutes a demand meta-model; it is a data abstraction of an independent information unit within the demand domain. A structured demand unit typically refers to a structured data object with unified, explicit, and atomic semantic slots, generated after the demand entity has undergone natural language processing and semantic parsing.

[0037] Step S30: Based on the demand entity and the structured demand unit, identify the logical relationship to obtain a set of demand logical relationships.

[0038] In one embodiment, the step of identifying logical relationships based on the demand entity and the structured demand units to obtain a set of demand logical relationships includes: identifying logical relationships based on the demand entity and the structured demand units to determine the identification results of each relationship; calculating the confidence level of the relationship identification results to obtain the confidence level corresponding to each relationship identification result; and obtaining a set of demand logical relationships based on the relationship identification results and the confidence levels corresponding to the relationship identification results.

[0039] The step of identifying logical relationships based on the demand entity and the structured demand unit, and determining the identification results of each relationship, includes: acquiring a preset domain knowledge rule base and text features in the unified demand text; generating candidate demand relationship pairs based on the structured demand unit using the text features; performing identification processing on the structured demand unit and the candidate demand relationship pairs to obtain relationship clues and semantic judgment results; performing relationship matching on the demand entity through the domain knowledge rule base to obtain intermediate results; and determining the identification results of each relationship based on the candidate demand relationship pairs, the relationship clues, the semantic judgment results, the intermediate results, and the domain knowledge rule base.

[0040] It should be noted that a domain knowledge rule base typically refers to an executable knowledge set formed by formally encoding domain expert experience, project specifications, and analysis strategies; a domain knowledge rule base is a fusion of domain factual knowledge and logical judgment rules. Candidate requirement relationship pairs are usually a subset of requirement pairs most likely to have logical relationships, selected from a massive number of possible requirement pairings through efficient and heuristic methods, and form the basis for subsequent in-depth analysis and accurate judgment.

[0041] In addition, the set of logical relationship requirements is usually a structured data set, in which each record precisely describes the logical relationship between two requirements after system identification and judgment; the set of logical relationship requirements is not only a simple list of relationships, but also a traceable set of relationship knowledge with judgment criteria and confidence. The logical relationship usually includes the following relationships: (1) inclusion relationship: the upper-level requirement includes several sub-requirements; (2) implication relationship: when one requirement is established, another requirement should be established at the same time; (3) conditional relationship: the corresponding function or constraint is triggered after a certain condition is met; (4) temporal relationship: there is a sequential execution order between multiple requirements; (5) derivation relationship: subsystem-level or component-level requirements are derived from system-level requirements; (6) dependency relationship: the implementation of one requirement depends on another requirement; (7) constraint relationship: performance, security, interface, environment and other constraints act on functional requirements; (8) conflict relationship: multiple requirements are inconsistent in terms of objects, parameters, states or boundaries; (9) duplication relationship: multiple requirements have high semantic similarity and there is duplicate description; (10) traceability relationship: there is a correlation between requirements, design, verification and test cases.

[0042] Specifically, a structured requirement unit should at least include "subject, action, object, condition, input, output, result state, and constraint parameters." Text features should at least include chapter hierarchy, entity co-occurrence, and semantic similarity. Based on text features, candidate requirement relationship pairs are automatically generated from the structured requirement units. Then, conditional sentences, temporal words, causal words, and constraint expressions can be identified through syntactic analysis (i.e., relationship clues are obtained). Semantic role labeling is used to determine whether the output, result, or state of one requirement constitutes the input, triggering condition, or constraint object of another requirement (i.e., semantic determination result). Next, the composition, calls, interfaces, and hierarchical relationships between entities are matched and verified using a domain knowledge rule base to obtain intermediate results. Finally, based on the candidate requirement relationship pairs, relationship clues, semantic determination results, intermediate results, and the domain knowledge rule base, the identification results of each relationship are determined. For example, if the output of the previous requirement corresponds to the output of the next requirement... If an input to a requirement is identified as a dependency, and the result of a previous requirement triggers the execution of a subsequent requirement, it is identified as a timing or conditional relationship (specifically, "when the temperature exceeds the threshold, the system should enter protection mode" can be identified as a conditional relationship; "the control module can only start after the system completes initialization" can be identified as a timing relationship); if a subsequent requirement refines the function or supplements the parameters of a previous requirement, it is identified as a derivation relationship; if two requirements give opposite states, ranges, or constraints to the same object, they are identified as conflicting relationships (specifically, "the controller response time does not exceed 100ms" and "the controller response time must not be less than 200ms" can be identified as parameter conflict relationships). Figure 3 As shown, Figure 3 Example diagram for identifying logical relationships.

[0043] After obtaining the identification results of each relationship, the confidence level corresponding to each identification result can be calculated, and the requirement number, relationship type, judgment basis and confidence level can be output to form a set of traceable requirement logical relationships.

[0044] This embodiment uses a combination of syntactic analysis, semantic role labeling, and domain knowledge rule base to automatically identify complex requirement logic relationships, thereby forming a traceable set of requirement logic relationships and improving the traceability of the topology graph acquisition process.

[0045] Step S40: Compare and score the set of logical relationships of requirements and the set of standard answers to obtain an evaluation score.

[0046] It should be noted that the evaluation metrics used in the comparative scoring mainly include accuracy, recall, and F1 score. Specifically, accuracy = number of correctly identified items / total number of items identified by the system; recall = number of correctly identified items / total number of manually labeled items; and F1 score = 2 × accuracy × recall / (accuracy + recall).

[0047] Specifically, the system compares the set of logical relationships of requirements with the standard answer item by item. During evaluation, entities or relationships correctly identified by the system are recorded as correct results, results output by the system but not found in the standard answer are recorded as misidentified results, and results found in the standard answer but not identified by the system are recorded as missed identification results. In addition, entity extraction requires that the entity boundaries and entity types are correct; relationship identification requires that the relationship type, associated entities, and relationship direction are correct; and evaluation scores are obtained for each evaluation indicator based on the evaluation process.

[0048] Step S50: When the evaluation score is greater than a preset threshold, a target demand topology map is generated based on the demand entity, the structured demand unit, and the set of demand logical relationships.

[0049] Understandably, when the evaluation score is greater than the preset threshold, it can be further confirmed by manual review that the set of requirement logic relationships can correctly support the construction of the requirement topology diagram, consistency check and conflict location before proceeding to the final step of generating the requirement topology diagram.

[0050] Alternatively, actual requirement documents from aerospace systems, automotive electronic control systems, or industrial control systems can be used for verification. The verification process includes: importing the original requirement document, performing semantic parsing, generating structured requirement entities, identifying requirement logical relationships, constructing a SysML requirement model, generating a requirement topology diagram, and performing consistency and conflict detection. Verification metrics include: 1. Demand Coverage Completeness Rate: The ratio of the number of automatically identified valid requirements to the number of manually labeled requirements, used to evaluate the system's ability to cover the original requirements.

[0051] 2. Entity extraction accuracy: The degree of consistency between the functional subjects, action objects, constraints, performance parameters, and verification conditions identified by the system and the results of manual annotation.

[0052] 3. Logical Relationship Accuracy: The degree of consistency between the automatically identified conditional relationships, temporal relationships, derivation relationships, dependency relationships, and conflict relationships and the results of manual annotation.

[0053] 4. Modeling time: Compare the time required to automatically generate SysML requirement models with the time required for manual modeling to evaluate the effectiveness of the modeling efficiency improvement.

[0054] 5. Distribution of conflict detection quantity and type: Statistical analysis of the number of duplicate requirements, parameter conflicts, boundary conflicts, ambiguous requirements, and logical non-closure issues detected by the system.

[0055] 6. Traceability Completeness Rate: Evaluate whether a complete mapping relationship is established between the original requirement text, structured requirement entities, SysML model elements, and topology graph nodes.

[0056] For easier understanding, please refer to Figure 4 , Figure 4 The following are case studies and indicator diagrams for verification. Through the above verification, the improvement of this method compared with manual modeling and traditional rule template methods in terms of requirement coverage, accuracy of logical relationship identification, modeling efficiency, conflict detection capability, and requirement traceability can be quantitatively evaluated.

[0057] This embodiment provides a method for generating a demand topology map. It involves acquiring a unified demand text, a preset domain terminology library, and a set of standard answers; performing semantic parsing based on the unified demand text, the preset domain terminology library, and a preset language model to obtain demand entities and structured demand units; identifying logical relationships based on the demand entities and the structured demand units to obtain a set of demand logical relationships; comparing and scoring the set of demand logical relationships with the set of standard answers to obtain an evaluation score; and generating a target demand topology map based on the demand entities, the structured demand units, and the set of demand logical relationships when the evaluation score exceeds a preset threshold. Through semantic parsing and logical relationship identification mechanisms, the method enables automated generation of demand topology maps, improving the efficiency of demand topology map generation.

[0058] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. The preset language model includes at least an entity recognition model and a semantic correction model. Based on this, please refer to Figure 5 Step S20 includes steps S201 to S205: Step S201: Based on the preset domain terminology library, map the different entity representations in the unified requirement text to unified entity representations.

[0059] It should be noted that the preset domain terminology library includes at least system name, subsystem name, functional module, interface object, performance index, security constraint, verification method, abbreviation and synonym mapping.

[0060] After determining the preset domain terminology library, different entity expressions in the unified requirement text can be mapped to unified entity expressions. For example, different expressions such as "controller," "ECU," and "electronic control unit" can be mapped to unified entities. In addition, for newly emerging terms, the system can perform candidate classification based on contextual co-occurrence relationships and semantic similarity, and write them into the preset domain terminology library after user confirmation to achieve updates with the preset domain terminology library.

[0061] Step S202: Input the unified requirement text into the entity recognition model for entity recognition to obtain the requirement entity.

[0062] It should be noted that entity recognition models can typically be BERT (Bidirectional Encoder Representations from Transformers) models. Demand entities are usually the smallest independent data units extracted from the original demand data source, forming the demand meta-model. A demand entity is an abstraction of a single natural language demand record, containing a unique identifier for the demand and the original natural language description text; it is the raw material for all subsequent structured processing and modeling.

[0063] In one embodiment, the step of inputting the unified requirement text into the entity recognition model for entity recognition to obtain the requirement entity includes: obtaining entity features of the unified requirement text; inputting the unified requirement text into the entity recognition model for entity recognition to obtain a context entity index and an entity pointer delivery; determining candidate entities and their corresponding confidence scores based on the entity features, the pointer delivery, and the context entity index; and obtaining the requirement entity based on the candidate entities and their corresponding confidence scores.

[0064] Specifically, a standardized requirement text is input into the BERT model. Entities such as systems, modules, interfaces, functions, parameters, and states that have appeared are recorded according to chapter, paragraph, and requirement number, forming a contextual entity index. The system then detects the referential expressions of these entities. When a referential expression is detected, the system determines the candidate range based on the type of the referent. For example, "the system" prioritizes matching system entities in the current chapter or a higher-level chapter; "the aforementioned module" prioritizes matching module entities that appeared before the current requirement; and "its output" prioritizes matching the most recently appearing function or interface entities with input / output attributes. After determining the candidate range, the system combines the contextual entity index and entity features (including at least entity location, chapter-level distance, entity type consistency, contextual co-occurrence, and semantic similarity) to determine candidate entities and their corresponding confidence scores. Candidate entities are ranked according to their confidence scores, and the candidate entity with the highest score exceeding the confidence threshold is selected as the requirement entity, replacing the original referential expression with a standardized entity name. Furthermore, if a candidate entity has insufficient scores or multiple similar candidates exist, the system marks the referential relationship as pending confirmation, prompting manual selection of the correct object.

[0065] Step S203: Input the unified requirement text into the semantic correction model for semantic correction to obtain the correction result.

[0066] It should be noted that semantic correction models are typically large language models. Semantic correction mainly refers to correcting semantic ambiguity and semantic contradictions in the requirement text.

[0067] In specific implementation, the step of inputting the unified requirement text into the semantic correction model for semantic correction to obtain the correction result includes: inputting the unified requirement text into the semantic correction model for semantic anomaly judgment to obtain anomaly judgment result; when the anomaly judgment result indicates that the unified requirement text has a semantic clarity problem, outputting a first correction suggestion based on the text content corresponding to the semantic clarity problem, wherein the first correction suggestion includes at least one of the following: prompting for supplementing specific parameters, prompting for supplementing a threshold range, prompting for supplementing applicable conditions, and prompting for supplementing verification standards; when the anomaly judgment result indicates that the unified requirement text has a logical consistency problem, outputting a second correction suggestion based on the text content corresponding to the logical consistency problem, wherein the second correction suggestion includes at least the conflicting requirement number, conflicting field, conflicting reason, and conflict correction suggestion; and obtaining the correction result based on the first correction suggestion and / or the second correction suggestion.

[0068] Specifically, the standardized requirement text is input into a large language model for semantic parsing and anomaly detection, yielding anomaly detection results. When the anomaly detection result indicates that the standardized requirement text has semantic clarity issues (in layman's terms, semantic ambiguity), such as expressions lacking clear boundaries like "as much as possible," "quickly," "appropriately," "when necessary," and "high precision," the system marks the requirement as an ambiguous requirement and prompts the user to supplement specific parameters, threshold ranges, usage conditions, or verification standards. When the anomaly detection result indicates that the standardized requirement text has logical consistency issues (in layman's terms, semantic contradictions or conflicts), for conflicting requirements, the system compares the requirement number, entity object, parameter fields, constraint range, and logical relationships to identify inconsistencies between different requirements in terms of function, state, timing, performance indicators, or boundary conditions, and outputs the conflicting requirement number, conflicting fields, conflict reasons, and correction suggestions. Finally, based on the first and / or second correction suggestions, a correction result can be obtained, which can then be used in subsequent topology graph generation steps.

[0069] Step S204: Based on the unified entity representation, the requirement entity, and the correction result, a structured requirement text is obtained.

[0070] Understandably, after semantic parsing is completed, the unified entity representation, requirement entity, and correction results can be further processed to obtain a structured requirement text.

[0071] In one possible implementation, the semantic parsing step can be performed within a semantic parsing technology architecture, referring to... Figure 6 , Figure 6 This is a schematic diagram of the semantic parsing technology architecture, which includes: inputting the input text into a pre-trained language model and a fine-tuning model, combining a domain terminology dictionary and a knowledge graph to perform entity recognition, relation extraction, attribute extraction and hierarchical recognition operations, and feeding back the data information during the semantic parsing process of the model to the domain terminology dictionary and knowledge graph for updating.

[0072] Step S205: Parse the structured requirement text to obtain structured requirement units.

[0073] The system parses and processes the structured requirement text, and finally parses it into structured requirement units containing "subject, action, object, condition, input, output, result status and constraint parameters".

[0074] This embodiment, based on the preset domain terminology library, maps different entity representations in the unified requirement text to unified entity representations; inputs the unified requirement text into the entity recognition model for entity recognition to obtain requirement entities; inputs the unified requirement text into the semantic correction model for semantic correction to obtain correction results; obtains structured requirement text based on the unified entity representations, the requirement entities, and the correction results; parses the structured requirement text to obtain structured requirement units; and achieves high-precision semantic parsing of complex requirement text by unifying entity representations, recognizing requirement entities, and semantic correction through semantic parsing methods.

[0075] Based on the first embodiment of this application, in the third embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 7 Step S50 includes steps S501 to S505: Step S501: When the evaluation score is greater than a preset threshold, a demand meta-model is established based on the demand entity, the structured demand unit, and the set of demand logical relationships.

[0076] It should be noted that a requirement meta-model is typically a formalized specification of the concepts, attributes, relationships, and constraints required for requirement modeling.

[0077] It should be understood that when the evaluation score exceeds a preset threshold, the system can combine the requirement entity, structured requirement unit, and set of requirement logical relationships to obtain the requirement parsing result. Then, the requirement number, requirement name, requirement text, requirement source, module, priority, constraint parameters, verification method, requirement status, and logical relationships with other requirements in the parsing result are processed into fields to form a unified data structure. For requirement documents from different sources, the system uses a field mapping table to convert the original fields into unified fields, thereby ensuring the consistency of multi-source requirement data. The standardized requirement meta-model can be represented in the following form: Requirement{ id: Request number name: Requirement Name text: Requirements body Source: Demand Source module: the module to which the module belongs priority: priority constraint: constraint parameters verifyMethod: Verification method status: Demand status relationList: A collection of logical relationships } Each requirement entity includes attributes such as requirement number, requirement name, requirement body, requirement source, module, priority, constraint parameters, verification method, and logical relationship, which are used to ensure data consistency and model uniformity among multiple requirements.

[0078] Step S502: Determine the system modeling language requirement elements based on the requirement meta-model and the set of requirement logical relationships.

[0079] It should be noted that the System Modeling Language (SML) requirement element is not plain text, but a modeled object with numbering, text, attributes, and relationships.

[0080] In one embodiment, the step of determining system modeling language requirement elements based on the requirement meta-model and the requirement logic relationship set includes: traversing each requirement entity in the requirement meta-model and creating requirement nodes corresponding to the requirement entities; writing field information from the requirement entities to the requirement nodes corresponding to the requirement entities to obtain first updated requirement nodes; traversing the relationship types in the requirement logic relationship set and determining the connection relationships between the first updated requirement nodes based on the relationship types; obtaining second updated requirement nodes based on the first updated requirement nodes and the connection relationships; grouping the second updated requirement nodes to obtain third updated requirement nodes; and determining system modeling language requirement elements based on the third updated requirement nodes.

[0081] Specifically, the process of generating requirement elements for a system modeling language can include: 1. Read the standardized requirement meta-model, traverse each requirement entity, and create a unique SysMLRequirement node for it.

[0082] 2. Write the field information from the requirement entity into the Requirement node attributes. The field information includes standard attributes such as id, name, and text, as well as extended attributes such as source, priority, module, constraint, and verifyMethod. The Requirement node with the field information written into it will be the first updated requirement node.

[0083] 3. Traverse the logical relationship set, call the corresponding SysML relationship template according to the relationship type in the logical relationship set, establish the connection relationship between the requirement nodes, and take the Requirement node after writing field information and establishing connection relationship as the second updated requirement node.

[0084] 4. Group the Requirement nodes according to their module, chapter level, or functional structure to form a package structure or requirement tree structure. The Requirement nodes that have undergone field information writing, connection relationship establishment, and grouping processing are used as the third updated requirement nodes.

[0085] 5. Finally, determine the system modeling language requirement elements based on the third updated requirement node: serialize and organize the information such as the identifier, attributes, structural hierarchy and optional layout coordinates of the generated third updated requirement node into a SysML model file, and output it in formats such as XMI, ReqIF, JSON, XML or CSV, which are the final SysML requirement elements.

[0086] In addition, a SysML requirement element is essentially a structured requirement node, which typically includes the following: < <requirement>>Requirement Name id=requirement number text=Requirement Body source = demand source module = the module to which it belongs priority constraint = constraint parameters verifyMethod = Verification Method For example, a natural language requirement is: REQ-001: The system should complete target recognition within 2 seconds, with an accuracy rate of no less than 95%. The converted SysML requirement elements can be represented as: < <requirement>Target recognition performance requirements id=REQ-001 The system should complete target recognition within 2 seconds, with an accuracy rate of no less than 95%.

[0087] Constraints: Response time ≤ 2s; Recognition accuracy ≥ 95% verifyMethod=test priority = high In addition, the relationships between SysML requirement elements typically include the following: •Requirement element; • DeriveReqt derivation relationship; • Satisfy satisfies the relation; •Verify the relationship; •Refine to refine relationships; • Trace relationships; • Dependency; • Containment is a relational term.

[0088] Each SysML requirement element contains attributes such as requirement number, requirement name, requirement body, requirement type, source document, module, priority, constraint parameters, validation method, status identifier, and association relationship.

[0089] This embodiment reads the standardized demand meta-model and traverses the demand entities, automatically creating demand nodes in batches and filling in their attributes. This avoids the tedious manual dragging and dropping of demand elements one by one, thereby improving the efficiency of topology graph generation.

[0090] Step S503: Generate an initial system modeling language model based on the requirement meta-model and the system modeling language requirement elements.

[0091] It should be noted that the initial system modeling language model is a SysML requirement model that has not yet undergone rule testing.

[0092] The system uses a requirement meta-model and system modeling language requirement elements to perform standardized field mapping and relation type mapping, transforming natural language requirements into an importable, traceable, verifiable, and maintainable SysML requirement model (i.e., the initial system modeling language model). The generated model not only contains individual requirement content but also includes the derivation, inclusion, verification, satisfaction, traceability, and conflict relationships between requirements.

[0093] Step S504: Perform rule detection on the initial system modeling language model, and use the initial system modeling language model that passes the rule detection as the target system modeling language model.

[0094] For example, for ease of understanding, refer to Figure 8 , Figure 8 This diagram illustrates consistency and conflict detection, including the rule base structure, detection dimensions, and output results and suggestions. Rule detection is typically performed using a rule base, which consists of at least three categories of rules. Each rule includes a rule number, applicable objects, triggering conditions, judgment logic, conflict type, and handling suggestions. Specifically, it includes: (1) The first category is SysML metamodel constraint rules, which are used to check whether the requirement elements are complete, whether the number is unique, whether the relationship type is legal, whether the requirement hierarchy is closed, and whether there are isolated nodes or illegal connections.

[0095] (2) The second category is domain engineering rules, which are used to check whether requirements violate domain constraints, system boundaries, interface definitions, security restrictions, or performance constraints. The functional objects, operating modes, interface parameters, and performance indicators in the requirements are compared with the domain rules. For example, if a function is only allowed to be triggered in manual mode, but another requirement requires it to be executed in automatic mode, the system identifies it as a system boundary conflict.

[0096] (3) The third category is semantic logic rules, which are used to detect duplicate requirements, ambiguous requirements, parameter conflicts, and logical contradictions. Detection is performed through semantic similarity, parameter normalization, negative word recognition, and conditional relationship judgment. For example, if two requirements constrain the same performance index of the same object, but the thresholds are different, the system marks them as parameter conflicts; if the semantic similarity of the two requirements exceeds the threshold, the system marks them as duplicate requirements.

[0097] After the detection is completed, if the initial system modeling language model fails the rule detection, the system outputs the problem requirement number, conflict type, triggering rule, conflict cause, and modification suggestions. It can also generate a requirement topology diagram, highlighting the corresponding nodes or relationship edges in the requirement topology diagram to assist users in reviewing and correcting. After the detection is completed, if the initial system modeling language model passes the rule detection, the initial system modeling language model that passes the rule detection will be used as the target system modeling language model.

[0098] In addition, the rule base can be built from predefined rules and expanded by combining historical requirement review records and machine learning methods. Specifically, the predefined rules are compiled from SysML modeling specifications, domain design specifications, interface control files, and security constraint files; typical conflict cases in historical review records are transformed into reusable rules after field extraction, conflict type labeling, and rule template summarization; for frequently occurring but not yet covered conflict types, the system forms candidate rules through cluster analysis and similar case mining, which are then added to the rule base after manual confirmation.

[0099] Step S505: Generate a target requirement topology diagram based on the target system modeling language model.

[0100] Understandably, after obtaining the target system modeling language model, a requirement topology graph can be further constructed. The system uses requirement entities as nodes and requirement logical relationships as edges to automatically form a requirement network structure. Furthermore, the topology graph layout algorithm also supports user customization. The system provides hierarchical layout, force-directed layout, and module clustering layout by default. Users can customize grouping and display according to requirement level, system module, functional domain, security level, verification method, or requirement status.

[0101] In one implementation, the system can simultaneously output a SysML requirement diagram, a (target) requirement topology diagram, and an output file, such as... Figure 9 As shown, Figure 9 The system output includes example diagrams, such as example diagrams of SysML requirement diagrams, example diagrams of requirement topology diagrams, and output files in different formats.

[0102] This embodiment establishes a requirement meta-model based on the requirement entity, the structured requirement unit, and the set of requirement logical relationships when the evaluation score is greater than a preset threshold; determines system modeling language requirement elements based on the requirement meta-model and the set of requirement logical relationships; generates an initial system modeling language model based on the requirement meta-model and the system modeling language requirement elements; performs rule detection on the initial system modeling language model and uses the initial system modeling language model that passes the rule detection as the target system modeling language model; generates a target requirement topology graph based on the target system modeling language model, and converts it into a computable, inferable, and modelable target requirement topology graph through unified semantic parsing and requirement meta-modeling, making the final target requirement topology graph more reliable.

[0103] For example, to help understand the implementation flow of the requirement topology graph generation method obtained by combining this embodiment with the above embodiment one, please refer to... Figure 10 , Figure 10 This paper presents a timeline diagram of a method for generating a requirement topology graph. Specifically, the system includes: a requirement text preprocessing module, a semantic parsing and structured expression module, a requirement logical relationship identification module, a SysML requirement model generation module, and a requirement topology graph construction and visualization module. The requirement text preprocessing module completes steps such as document parsing, sentence segmentation, paragraphing, and number recognition; the semantic parsing and structured expression module completes steps such as entity recognition, relationship processing, and implicit requirement identification; the requirement logical relationship identification module completes the relationship type identification process; the SysML requirement model generation module generates a SysML requirement model that conforms to the model specification; and the requirement topology graph construction and visualization module constructs a requirement topology graph containing edges and nodes and supporting user-defined layouts, thereby completing the overall process of generating the requirement topology graph.

[0104] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the topology graph generation method required by this application. Any simple transformations based on this technical concept are within the protection scope of this application.

[0105] This application also provides a demand topology graph generation apparatus; please refer to... Figure 11 The demand topology graph generation device includes: Module 10 is used to obtain standardized requirement text, a preset domain terminology library, and a set of standard answers.

[0106] Processing module 20 is used to perform semantic parsing based on the unified requirement text, the preset domain terminology library, and the preset language model to obtain requirement entities and structured requirement units.

[0107] The processing module 20 is also used to identify logical relationships based on the demand unit and the structured demand unit to obtain a set of demand logical relationships.

[0108] The processing module 20 is also used to compare and score the set of logical relationships of requirements and the set of standard answers to obtain an evaluation score.

[0109] The generation module 30 is used to generate a target demand topology map based on the demand entity, the structured demand unit, and the set of demand logical relationships when the evaluation score is greater than a preset threshold.

[0110] The demand topology graph generation apparatus provided in this application, employing the demand topology graph generation method described in the above embodiments, can solve the technical problem of low generation efficiency caused by the reliance on manual operation for semantic understanding and logical recognition in the current demand topology graph generation process. Compared with the prior art, the beneficial effects of the demand topology graph generation apparatus provided in this application are the same as those of the demand topology graph generation method provided in the above embodiments, and other technical features in the demand topology graph generation apparatus are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0111] In one embodiment, the processing module 20 is further configured to identify logical relationships based on the demand entity and the structured demand unit, determine the identification results of each relationship, calculate the confidence level of the relationship identification results, obtain the confidence level corresponding to each relationship identification result, and obtain a set of demand logical relationships based on the relationship identification results and the confidence levels corresponding to the relationship identification results.

[0112] In one embodiment, the processing module 20 is further configured to acquire a preset domain knowledge rule base and text features in the unified requirement text; generate candidate requirement relationship pairs based on the text features and the structured requirement units; perform identification processing on the structured requirement units and the candidate requirement relationship pairs to obtain relationship clues and semantic judgment results; perform relationship matching on the requirement entities through the domain knowledge rule base to obtain intermediate results; and determine the relationship identification results based on the candidate requirement relationship pairs, the relationship clues, the semantic judgment results, the intermediate results, and the domain knowledge rule base.

[0113] In one embodiment, the processing module 20 is further configured to: map different entity representations in the unified requirement text to unified entity representations based on the preset domain terminology library; input the unified requirement text into the entity recognition model for entity recognition to obtain requirement entities; input the unified requirement text into the semantic correction model for semantic correction to obtain correction results; obtain structured requirement text based on the unified entity representations, the requirement entities, and the correction results; and parse the structured requirement text to obtain structured requirement units.

[0114] In one embodiment, the processing module 20 is further configured to: acquire entity features of the unified requirement text; input the unified requirement text into the entity recognition model for entity recognition to obtain a context entity index and an entity instruction delivery; determine candidate entities and their corresponding confidence scores based on the entity features, the instruction delivery, and the context entity index; and obtain the requirement entity based on the candidate entities and their corresponding confidence scores.

[0115] In one embodiment, the processing module 20 is further configured to input the unified requirement text into the semantic correction model for semantic anomaly judgment to obtain an anomaly judgment result; when the anomaly judgment result indicates that the unified requirement text has a semantic clarity problem, output a first correction suggestion based on the text content corresponding to the semantic clarity problem, wherein the first correction suggestion includes at least one of the following: prompting for supplementing specific parameters, prompting for supplementing a threshold range, prompting for supplementing applicable conditions, and prompting for supplementing verification standards; when the anomaly judgment result indicates that the unified requirement text has a logical consistency problem, output a second correction suggestion based on the text content corresponding to the logical consistency problem, wherein the second correction suggestion includes at least the conflicting requirement number, conflicting field, conflicting reason, and conflict correction suggestion; and obtain a correction result based on the first correction suggestion and / or the second correction suggestion.

[0116] In one embodiment, the generation module 30 is further configured to: establish a requirement meta-model based on the requirement entity, the structured requirement unit, and the requirement logical relationship set when the evaluation score is greater than a preset threshold; determine system modeling language requirement elements according to the requirement meta-model and the requirement logical relationship set; generate an initial system modeling language model based on the requirement meta-model and the system modeling language requirement elements; perform rule detection on the initial system modeling language model, and use the initial system modeling language model that passes the rule detection as the target system modeling language model; and generate a target requirement topology diagram according to the target system modeling language model.

[0117] In one embodiment, the generation module 30 is further configured to: traverse each requirement entity in the requirement meta-model and create requirement nodes corresponding to the requirement entities; write field information from the requirement entities to the requirement nodes corresponding to the requirement entities to obtain first updated requirement nodes; traverse the relationship types in the requirement logic relationship set and determine the connection relationships between the first updated requirement nodes based on the relationship types; obtain second updated requirement nodes based on the first updated requirement nodes and the connection relationships; group the second updated requirement nodes to obtain third updated requirement nodes; and determine system modeling language requirement elements based on the third updated requirement nodes.

[0118] This application provides a demand topology graph generation device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the demand topology graph generation method in Embodiment 1 above.

[0119] The following is for reference. Figure 12 This document illustrates a structural schematic diagram of a demand topology graph generation device suitable for implementing embodiments of this application. The demand topology graph generation device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), tablets, PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 12 The demand topology graph generation device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0120] like Figure 12 As shown, the demand topology map generation device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in ROM (Read Only Memory) 1002 or a program loaded from storage device 1003 into RAM (Random Access Memory) 1004. RAM 1004 also stores various programs and data required for the operation of the demand topology map generation device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via bus 1005. Input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, LCDs (Liquid Crystal Displays), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the demand topology graph generation device to communicate wirelessly or wiredly with other devices to exchange data. Although a demand topology graph generation device with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0121] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0122] The demand topology graph generation device provided in this application, employing the demand topology graph generation method described in the above embodiments, can solve the technical problem of low generation efficiency caused by the reliance on manual operation for semantic understanding and logical recognition in the current demand topology graph generation process. Compared with the prior art, the beneficial effects of the demand topology graph generation device provided in this application are the same as those of the demand topology graph generation method provided in the above embodiments, and other technical features in this demand topology graph generation device are the same as those disclosed in the previous embodiment method, and will not be repeated here.

[0123] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0124] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should 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.

[0125] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the demand topology graph generation method in the above embodiments.

[0126] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, RAM (Random Access Memory), ROM (Read Only Memory), EPROM (Erasable Programmable Read Only Memory or Flash Memory), optical fibers, CD-ROM (CD-Read Only Memory), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0127] The aforementioned computer-readable storage medium may be included in the demand topology graph generation device; or it may exist independently and not assembled into the demand topology graph generation device.

[0128] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by the demand topology graph generation device, the demand topology graph generation device: acquires a unified demand text, a preset domain terminology library, and a set of standard answers; performs semantic parsing based on the unified demand text, the preset domain terminology library, and a preset language model to obtain demand entities and structured demand units; identifies logical relationships based on the demand entities and the structured demand units to obtain a set of demand logical relationships; compares and scores the set of demand logical relationships with the set of standard answers to obtain an evaluation score; and when the evaluation score is greater than a preset threshold, generates a target demand topology graph based on the demand entities, the structured demand units, and the set of demand logical relationships.

[0129] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as "C" or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including LAN (Local Area Network) or WAN (Wide Area Network)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0130] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0131] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0132] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described demand topology graph generation method. This solves the technical problem of low generation efficiency caused by the reliance on manual operation for semantic understanding and logical recognition in the current demand topology graph generation process. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the demand topology graph generation method provided in the above embodiments, and will not be repeated here.

[0133] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the demand topology graph generation method described above.

[0134] The computer program product provided in this application can solve the technical problem of low generation efficiency caused by the reliance on manual operation for semantic understanding and logical recognition in the current demand topology graph generation process. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the demand topology graph generation method provided in the above embodiments, and will not be repeated here.

[0135] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.< / requirement> < / requirement>

Claims

1. A method for generating a demand topology graph, characterized in that, The method for generating the demand topology graph includes: Obtain a unified requirement text, a pre-defined domain terminology library, and a standard answer set; Semantic parsing is performed based on the unified requirement text, the preset domain terminology library, and the preset language model to obtain requirement entities and structured requirement units; Logical relationships are identified based on the demand entities and the structured demand units to obtain a set of demand logical relationships; The set of logical relationships of the requirements and the set of standard answers are compared and scored to obtain an evaluation score; When the evaluation score is greater than a preset threshold, a target demand topology diagram is generated based on the demand entity, the structured demand unit, and the set of demand logical relationships.

2. The method as described in claim 1, characterized in that, The step of identifying logical relationships based on the demand entity and the structured demand unit to obtain a set of demand logical relationships includes: Logical relationships are identified based on the demand entities and the structured demand units, and the identification results of each relationship are determined. Calculate the confidence level of the relationship identification results to obtain the confidence level corresponding to each relationship identification result; The set of logical relationships of the requirements is obtained based on the relationship identification results and the confidence levels corresponding to the relationship identification results.

3. The method as described in claim 2, characterized in that, The step of identifying logical relationships based on the demand entity and the structured demand unit, and determining the identification results of each relationship, includes: Obtain the preset domain knowledge rule base and the text features in the unified requirement text; Based on the text features, candidate demand relationship pairs are generated based on the structured demand units; The structured demand units and the candidate demand relationship pairs are identified and processed to obtain relationship clues and semantic determination results; Intermediate results are obtained by performing relationship matching on the required entities using the domain knowledge rule base; The identification results of each relation are determined based on the candidate demand relation pairs, the relation clues, the semantic determination results, the intermediate results, and the domain knowledge rule base.

4. The method as described in claim 1, characterized in that, The preset language model includes at least an entity recognition model and a semantic correction model. The step of performing semantic parsing based on the unified requirement text, the preset domain terminology library, and the preset language model to obtain requirement entities and structured requirement units includes: Based on the preset domain terminology library, the different entity representations in the unified requirement text are mapped to unified entity representations; The unified requirement text is input into the entity recognition model for entity recognition to obtain the requirement entity; The unified requirement text is input into the semantic correction model for semantic correction, and the correction result is obtained. A structured requirement text is obtained based on the unified entity representation, the requirement entity, and the correction result. The structured requirement text is parsed to obtain structured requirement units.

5. The method as described in claim 4, characterized in that, The step of inputting the unified requirement text into the entity recognition model for entity recognition to obtain the requirement entity includes: Obtain the entity features of the unified requirement text; The unified requirement text is input into the entity recognition model for entity recognition, and the context entity index and entity pointer expression are obtained. Based on the entity features, the index expression, and the context entity index, a candidate entity and its corresponding confidence score are determined. The required entity is obtained based on the candidate entity and the confidence score corresponding to the candidate entity.

6. The method as described in claim 4, characterized in that, The step of inputting the unified requirement text into the semantic correction model for semantic correction and obtaining the correction result includes: The unified requirement text is input into the semantic correction model for semantic anomaly detection to obtain the anomaly detection result. When the anomaly determination result indicates that the unified requirement text has a semantic clarity problem, a first correction suggestion is output based on the text content corresponding to the semantic clarity problem. The first correction suggestion includes at least one of the following: prompting for supplementation of specific parameters, prompting for supplementation of threshold range, prompting for supplementation of applicable conditions, and prompting for supplementation of verification standards. When the anomaly determination result indicates that there is a logical consistency problem in the unified requirement text, a second correction suggestion is output based on the text content corresponding to the logical consistency problem. The second correction suggestion includes at least the conflicting requirement number, conflicting field, conflicting reason, and conflict correction suggestion. The correction result is obtained based on the first correction suggestion and / or the second correction suggestion.

7. The method as described in claim 1, characterized in that, The step of generating a target demand topology map based on the demand entity, the structured demand unit, and the set of demand logical relationships when the evaluation score is greater than a preset threshold includes: When the evaluation score is greater than a preset threshold, a demand meta-model is established based on the demand entity, the structured demand unit, and the set of demand logical relationships. The system modeling language requirement elements are determined based on the requirement meta-model and the set of requirement logical relationships. An initial system modeling language model is generated based on the aforementioned requirement meta-model and the requirement elements of the system modeling language. The initial system modeling language model is subjected to rule detection, and the initial system modeling language model that passes the rule detection is used as the target system modeling language model; Generate a target requirement topology diagram based on the target system modeling language model.

8. The method as described in claim 7, characterized in that, The step of determining the system modeling language requirement elements based on the requirement meta-model and the set of requirement logical relationships includes: Iterate through each demand entity in the demand meta-model and create the demand node corresponding to the demand entity; Write the field information in the demand entity to the demand node corresponding to the demand entity to obtain the first updated demand node; Traverse the relationship types in the set of requirement logical relationships, and determine the connection relationships between the first updated requirement nodes based on the relationship types; Based on the first updated demand node and the connection relationship, the second updated demand node is obtained; The second updated demand nodes are grouped to obtain the third updated demand nodes; The system modeling language requirement elements are determined based on the third updated requirement node.

9. A demand topology graph generation device, characterized in that, The demand topology graph generation device includes: The acquisition module is used to acquire standardized requirement text, a preset domain terminology library, and a set of standard answers. The processing module is used to perform semantic parsing based on the unified requirement text, the preset domain terminology library, and the preset language model to obtain requirement entities and structured requirement units. The processing module is also used to identify logical relationships based on the demand unit and the structured demand unit to obtain a set of demand logical relationships; The processing module is also used to compare and score the set of logical relationships of requirements and the set of standard answers to obtain an evaluation score; The generation module is used to generate a target demand topology map based on the demand entity, the structured demand unit, and the set of demand logical relationships when the evaluation score is greater than a preset threshold.

10. A demand topology graph generation device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the demand topology graph generation method as described in any one of claims 1 to 8.