Element association verification and change tracing method for complex architecture design

Abstract

The application relates to the technical field of system engineering and artificial intelligence, and discloses a factor correlation verification and change tracing method for complex system architecture design, which comprises the following steps: constructing a design factor standard dictionary and a domain specification constraint library; automatically extracting design factors and structurally representing the design factors in a five-tuple mode; constructing a system architecture design knowledge graph; intelligently correlating and matching the design factors; automatically detecting architecture correlation conflicts and performing consistency verification; and tracing and deducing the influence of design changes on the whole link. The application can realize whole-link tracing and iterative management and control of system architecture design changes, and guarantee the rigor, traceability and maintainability of complex system architecture design.

Description

Technical Field

[0001] This invention relates to the interdisciplinary field of systems engineering and artificial intelligence, and in particular to a method for element association verification and change traceability for complex system architecture design. Background Technology

[0002] Complex system architecture design involves the overall planning and collaborative design of multi-level and multi-type systems, components, interfaces, functions, constraints, performance indicators, and business interaction behaviors. It is widely used in engineering construction in fields such as large-scale information systems, industrial internet, smart cities, and rail transit. As the scale of systems continues to expand and the complexity of functions continues to increase, the number of design elements covered by the architecture design grows exponentially, and the relationships between these elements become increasingly intricate, placing higher demands on the standardization, accuracy, and maintainability of the architecture design.

[0003] Traditional system architecture design relies heavily on manual experience and offline documentation, which has significant shortcomings in practical engineering applications. First, the sources of various design elements are scattered, and their expressions are heterogeneous, lacking unified terminology and structured expression. Inconsistencies in terminology and format exist between design documents produced by different teams and at different stages, hindering machine-aided parsing and intelligent calculation, and impeding cross-team collaboration and communication. Second, the relationships between elements are mainly maintained manually, line by line, which is labor-intensive, inefficient, and prone to omissions and mismatches, especially in complex systems with a large number of elements, where the quality and efficiency of manual work are difficult to guarantee. Third, there is a lack of professional architectural conflict detection mechanisms, making it impossible to fully identify architectural defects such as circular dependencies, missing dependencies, constraint violations, performance mismatches, and topology violations. Many hidden logical flaws often only surface during later integration or testing phases, leading to high rectification costs and even impacting the overall project schedule. Fourth, the architecture design is frequently iterated and changed, and the existing methods are difficult to trace the impact of the entire chain. The scope of the impact of the changes can only be assessed by human experience. The accuracy and comprehensiveness of the assessment results depend on the personal experience of the designers, and the incomplete assessment can easily lead to subsequent chain design risks.

[0004] While existing publicly available technologies have conducted research on individual technical aspects such as element extraction, knowledge graph construction, and association matching, a complete technical solution covering standardized modeling of design elements, intelligent association matching, architecture conflict verification, and full-link traceability of design changes has not yet been formed. This cannot meet the actual engineering needs of standardization, automation, and traceability in complex system architecture design. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a method for element association verification and change traceability in complex system architecture design, which can meet the practical engineering needs for standardization, automation, and traceability in complex system architecture design.

[0006] The technical solution adopted in this invention is as follows: A method for element association verification and change traceability in complex system architecture design includes: The standardization and organization of multiple design elements forms a standard dictionary of design elements, integrates various domain architecture constraint specifications to build a domain specification constraint library, and extracts design elements from multiple source documents and encapsulates them into a five-tuple format; the standardization and organization includes classification and sorting, terminology normalization, and unified definition of attribute templates. A knowledge graph for system architecture design is constructed using five-tuple design elements as nodes and architectural logical relationships as directed edges. Multiple types of core relationships are defined and attributes and weights are configured. Based on the standard dictionary of the design elements and the domain specification constraint library, a multi-dimensional weighted similarity model is used to complete element association matching and confidence determination. A rule engine combining a traversal algorithm with the domain specification constraint library is used to compare the entire path of the knowledge graph to identify and locate multiple types of architectural conflicts; when design elements change, the changes propagate forward and backward along the core relationships to distinguish the impact level and quantitatively evaluate the fluctuation of architectural performance.

[0007] Furthermore, the standardization and regularization of multiple design elements includes: For various design elements, synonym sorting and unified mapping are performed to determine a unique standard name for each design element; standardized attribute templates are developed for different categories of design elements, clarifying the fixed attribute fields and data format specifications for each type of element, and design elements of the same category are all structured using the corresponding unified attribute template; By combining rule matching and semantic parsing, entity elements and associated attributes are identified from multi-source design documents. Rule matching relies on domain keywords and element feature rules to match the explicitly stated design entities and their corresponding attributes, while semantic parsing combines contextual semantic association to mine the implicit element association information in the text. The extracted elements undergo preliminary validity checks, verifying field completeness, format standardization, and element uniqueness, and eliminating invalid element entries with missing fields, incorrect formats, or duplicates.

[0008] Furthermore, the quintuple format includes: The unique identifier field is used to assign a globally unique identification code to each design element; The feature name field is used to record the standard naming of features after terminology normalization; The feature type field is used to identify the specific category to which a feature belongs. The attribute collection field is used to store the various attribute contents corresponding to the features. The attributes of each feature itself follow the attribute template specifications of the corresponding category. The Feature Source field records the original document or specification source from which the feature originated.

[0009] Furthermore, the domain specification constraint library includes: Hard constraints are rigid rules that the architecture design must meet. These hard constraints include topology specifications, parameter threshold boundaries, and legal association requirements. Soft recommendation constraints serve as reference rules for industry-wide best practices and historical engineering adaptation experience.

[0010] Furthermore, the multiple types of core relationships include: Compositional relationships represent the hierarchical and subordinate relationships between the whole and its parts; Dependency relationships represent the dependent associations that enable the implementation of functions; that is, functions cannot be implemented normally if the corresponding objects are missing. Constraint relationships characterize the parameter-limited associations between constraint elements and entity elements and functional elements; The performance support relationship characterizes the supporting relationship between functional elements and performance indicators. Functional performance directly determines the effectiveness of the corresponding performance indicators. Interaction relationships represent the bidirectional or unidirectional transmission and exchange of data and instructions between elements; Evolutionary influence relationships characterize the transmission correlation of element changes and are used for tracing design change links.

[0011] Furthermore, the multi-dimensional weighted similarity model includes: The feature type matching degree calculation module is used to determine whether the types of the features to be matched are consistent; The attribute set similarity calculation module is used to calculate the ratio of the intersection to the union of attribute sets; The domain specification rule matching degree calculation module is used to calculate the degree to which elements conform to domain constraint rules.

[0012] Furthermore, the completion of element association matching and confidence determination includes: Based on the similarity calculation results, historical association frequency parameters and constraint compliance parameters are introduced to correct the calculation results and obtain the element association confidence. A multi-level confidence threshold is set to divide the confidence interval into high-confidence, medium-confidence, and low-confidence intervals; the matching results in the high-confidence interval automatically establish stable element associations; the matching results in the medium-confidence interval generate candidate associations and push them; the matching results in the low-confidence interval are judged as invalid associations and directly filtered. The finalized relationships between elements are updated in real time to the system architecture design knowledge graph.

[0013] Furthermore, the identification and location of multiple types of architectural conflicts includes: Starting with the top-level elements of the system as the traversal starting node, it covers all sub-elements and their relationships layer by layer; it automatically identifies and locates five types of architectural defects: circular dependency conflicts, missing dependency conflicts, constraint violation conflicts, performance mismatch conflicts, and topology violation conflicts. Automatically label conflict locations, conflict types, and judgment criteria, and output corresponding correction suggestions based on the domain specification constraint library; after the rectification is completed and confirmed, trigger a second full-domain review, iteratively verifying until the entire domain architecture meets the domain specification constraint requirements.

[0014] Furthermore, the forward propagation and reverse tracing along the core relationship includes: Forward propagation starts with the changed design element as the starting node and extends the change propagation range layer by layer along the relationship link; backward tracing starts with the affected element as the base point and traces back to locate the source of the change; the entire domain is traversed through the bidirectional link of forward propagation and backward tracing. The traversal process relies on all the relationship types of the knowledge graph and performs a full path traversal along the architectural dependency links; when the parameters, functional configurations, or relationships of the design elements within the system architecture are added, modified, or deleted, the tracing process is triggered.

[0015] Furthermore, the distinction of impact levels and the quantification of architectural performance fluctuations include: The factors affected by the change are divided into three levels: directly affected factors, indirectly affected factors, and disturbed performance indicators. The directly affected factors are those directly related to the changed factor. The indirectly affected factors are those affected after being transmitted through multiple layers of correlation. The disturbed performance indicators are performance indicators that fluctuate due to the change of factors. By combining the correlation hierarchy and coupling to quantify and evaluate the fluctuation range of the architecture topology and overall performance, a visualized change impact analysis report is generated. The report includes change information, transmission path, impact hierarchy, performance fluctuation data, and optimization suggestions. The updated element parameters, relationships, and performance evaluation results are synchronously and incrementally updated to the system architecture design knowledge graph, forming a closed-loop iterative management and control.

[0016] The beneficial effects of this invention are as follows: 1. This invention integrates scattered and heterogeneous architectural design information into a unified computable semantic network through an integrated technical architecture that combines standardized modeling of design elements, knowledge graph construction, intelligent association matching, architectural conflict detection, and change tracing. This enables a systematic processing of the entire process of complex system architecture design, from element extraction to conflict verification and change tracing. It can improve the standardization and consistency of architectural design to a certain extent, reduce the workload of manual review, and reduce the risk of architectural defects being exposed later.

[0017] 2. This invention, through the establishment of a standard dictionary of design elements and a five-tuple structured format, unifies the classification and terminology of various design elements in complex system architectures, and formulates standardized attribute templates. This effectively solves the problems of heterogeneous element descriptions and inconsistent formats in multi-source design documents, providing a unified data foundation for subsequent machine automatic parsing and intelligent calculation. The construction of the domain specification constraint library integrates industry design standards and engineering constraint rules in a structured manner, providing a unified judgment benchmark for the entire process and helping to ensure the consistency of judgment standards at each stage.

[0018] 3. This invention employs a multi-dimensional weighted similarity model for intelligent element association matching, comprehensively considering factors such as element type, attribute characteristics, and rule constraints. It also incorporates historical association frequency and constraint compliance parameters for correction, enabling a more accurate determination of the degree of association between elements compared to single-dimensional matching methods. Combined with a three-level confidence rating mechanism, it ensures the efficiency of automated processing of high-confidence associations while guaranteeing the accuracy of medium-confidence associations through manual verification, thus achieving a balance between efficiency and accuracy. Furthermore, the conflict detection method, based on a knowledge graph-based full-path traversal and a rule engine, can comprehensively identify various architectural defects, helping to discover potential problems in the early design stages and reducing later rectification costs.

[0019] 4. This invention relies on the bidirectional traceability capability of knowledge graphs. When design elements change, the entire path is traversed through a bidirectional link of forward propagation and backward tracing, which can comprehensively assess the transmission scope and impact of the change. By distinguishing between three levels—direct impact, indirect impact, and performance disturbance—and combining the coupling degree of the associated levels for quantitative evaluation, it can provide designers with a clearer view of the impact of changes, assisting in design decisions. The incremental update and closed-loop iteration mechanism of change results helps ensure the timeliness of knowledge graph data and achieve continuous iterative optimization of architecture design. Attached Figure Description

[0020] Figure 1 This is a flowchart of a method for element association verification and change tracing for complex system architecture design, according to Embodiment 2 of the present invention.

[0021] Figure 2This is a schematic diagram of the knowledge graph structure of the system architecture design in Embodiment 2 of the present invention. Detailed Implementation

[0022] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments are now described. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention; that is, the described embodiments are only a part of the embodiments of the invention, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0023] Example 1 This embodiment provides a method for element association verification and change traceability for complex system architecture design, including: The standardization and organization of multiple design elements forms a standard dictionary of design elements, integrates various domain architecture constraint specifications to build a domain specification constraint library, and extracts design elements from multiple source documents and encapsulates them into a five-tuple format; the standardization and organization includes classification and sorting, terminology normalization, and unified definition of attribute templates. A knowledge graph for system architecture design is constructed using five-tuple design elements as nodes and architectural logical relationships as directed edges. Multiple types of core relationships are defined and attributes and weights are configured. Based on the standard dictionary of the design elements and the domain specification constraint library, a multi-dimensional weighted similarity model is used to complete element association matching and confidence determination. A rule engine combining a traversal algorithm with the domain specification constraint library is used to compare the entire path of the knowledge graph to identify and locate multiple types of architectural conflicts; when design elements change, the changes propagate forward and backward along the core relationships to distinguish the impact level and quantitatively evaluate the fluctuation of architectural performance.

[0024] Specifically, the method of this embodiment can be implemented by the following steps: S1: Construct a standard dictionary of design elements and a domain specification constraint library The design of complex system architectures is categorized and organized into seven major categories: system elements, component elements, interface elements, functional elements, constraint elements, performance elements, and interaction elements. Terminology is standardized and attribute templates are uniformly defined to form a standard dictionary of design elements. At the same time, industry design standards, engineering constraint rules, and architecture topology compliance requirements are integrated to build a domain-specific constraint library, providing a unified benchmark constraint basis for subsequent design element extraction, intelligent association matching, architecture conflict verification, and change traceability simulation.

[0025] S2: Automatic extraction of design elements and structured representation of quintuples A joint extraction algorithm combining rule matching and semantic parsing is adopted to automatically identify entity elements and associated attributes from requirements documents, design specifications, architecture description documents, and parameter configuration tables. Synonym mapping, unit normalization, format regularization, and duplicate element removal are performed in sequence. All design elements are uniformly encapsulated into a standardized five-tuple structured format. The five-tuple consists of a unique identifier ID, element name, element type, attribute set, and element source, realizing the computable semantic expression of multi-source heterogeneous design elements.

[0026] S3: Building a Knowledge Graph for System Architecture Design Using the seven standardized design elements as graph nodes and the internal logical connections of the architecture as directed relation edges, a knowledge graph for system architecture design is constructed. Six core relational relationships are defined: compositional relationships, dependency relationships, constraint relationships, performance support relationships, interaction relationships, and evolutionary influence relationships. Inherent attributes, weight coefficients, and initial confidence levels are configured for graph nodes and relation edges to construct a traversable, computable, and deductive system semantic network carrier.

[0027] S4: Intelligent Association and Matching of Design Elements We construct a weighted attribute similarity calculation model for complex system architecture design scenarios. This model integrates element type matching degree, attribute set Jaccard similarity, and domain specification rule matching degree for multi-dimensional comprehensive calculation. We also combine historical association frequency score and constraint compliance score to correct the element association confidence. We set hierarchical confidence judgment thresholds and complete automatic element association, candidate association push, and invalid match filtering according to the threshold levels, thereby realizing intelligent semi-automatic rapid construction of system architecture topology.

[0028] S5: Automatic Detection and Consistency Verification of Architectural Association Conflicts By employing a depth-first traversal algorithm combined with a domain rule engine, the entire knowledge graph is traversed and its edges are compared against all rules. This enables the automatic identification and location of five types of architectural defects: circular dependency conflicts, missing dependency conflicts, constraint violation conflicts, performance mismatch conflicts, and topology violation conflicts. The system automatically marks the conflict location and type, matches the domain specification constraint library to provide corresponding correction suggestions, and initiates a second review and iteration after manual confirmation of the correction until the entire architecture meets the domain specification constraint requirements.

[0029] S6: End-to-end traceability and simulation of the impact of design changes When the parameters, functional configurations, or relationships of design elements within the system architecture are added, modified, or deleted, the system relies on the six types of relationships in the knowledge graph and uses forward propagation and backward tracing algorithms along the architecture dependency links to traverse the entire path. It automatically distinguishes directly affected elements, indirectly affected elements, and disturbed performance indicators, quantitatively assesses the fluctuation range of the architecture topology and overall performance, and generates a visualized change impact analysis report, realizing full-link tracing and quantitative assessment of complex system architecture design changes.

[0030] It should be noted that this method standardizes and classifies complex system architecture design elements and unifies attribute template specifications, enabling unified definitions and standardized expressions for various design elements, thus improving the standardization and consistency of system architecture design. Based on the constructed system architecture design knowledge graph and preset core relationships, combined with weighted similarity calculation, intelligent element association matching is achieved, reducing reliance on manual sorting and improving the efficiency of architecture element association configuration. By conducting association conflict detection and consistency verification through full path traversal combined with a domain rule engine, multiple types of system architecture design defects can be identified in advance, reducing design rework and subsequent rectification costs. This method enables full-link traceability and iterative control of system architecture design changes, ensuring the rigor, traceability, and maintainability of complex system architecture design.

[0031] Example 2 This embodiment is based on embodiment 1: like Figure 1 As shown, this embodiment provides a method for element association verification and change traceability for complex system architecture design, including the following steps: S1: Construct a standard dictionary of design elements and a domain specification constraint library This step categorizes and organizes the seven major categories of elements covered in complex system architecture design—system elements, component elements, interface elements, functional elements, constraint elements, performance elements, and interaction elements—standardizes terminology, and defines attribute templates in a unified manner, forming a standard dictionary of design elements.

[0032] Among them, terminology standardization: In response to the problems of abbreviations, common names, and multiple expressions of various design elements, the synonyms are sorted out and uniformly mapped, a unique standard name is determined for each design element, the problem of the same thing having different names and the same name having different meanings is avoided, and the terminology of all design elements in the system architecture is unified and the expression is standardized.

[0033] Unified definition of attribute templates: Standardized attribute templates are developed for each of the seven categories of design elements, clarifying the fixed attribute fields and data format specifications for each type of element; for design elements of the same category, the corresponding unified attribute template is used to complete the structured regularization, ensuring that the attribute structure of similar elements is consistent and can be standardized by machines for parsing and calculation.

[0034] The main fields of the attribute templates for the seven categories of design elements are as follows: System element attribute template: Element number, element name, platform category, functional positioning, basic features, data type, value range; Component feature attribute template: Feature ID, Feature Name, Component Type, System, Component Characteristics, Data Type, Value Range; Interface feature attribute template: Feature ID, Feature Name, Interface Type, Protocol Type, Interface Characteristics, Data Type, Value Range; Functional Element Attribute Template: Element Number, Element Name, Function Category, Dependency Relationship, Implementation Condition, Data Type, Value Range Constraint Feature Attribute Template: Feature ID, Feature Name, Constraint Type, Applicable Object, Threshold Specification, Data Type, Value Range; Performance Element Attribute Template: Element Number, Element Name, Indicator Category, Statistical Scope, Target Requirements, Data Type, Value Range; Interactive element attribute template: Element number, Element name, Interactive subject, Interactive object, Interactive mode, Data type, Value range.

[0035] Examples of typical design elements included in the standard dictionary of design elements: System elements: reconnaissance platform, communication relay platform, command and control platform; Component elements: environmental perception component, data link communication component, task control component, and energy security component; Interface elements: detection data interaction interface, control command issuance interface, and device status reporting interface; Functional elements: autonomous target detection, multi-machine cooperative positioning, communication relay forwarding, and task command execution; Constraints: Platform operating distance constraint, communication bandwidth constraint, transmission delay constraint, and operation permission constraint; Performance factors: regional detection coverage, network communication connectivity, mission response latency, and overall mission success rate; Interactive elements: Detection data is reported level by level, control instructions are issued layer by layer, status information is shared synchronously, and cross-platform relay forwarding is implemented.

[0036] Based on a standard dictionary of design elements, a domain-specific constraint library is simultaneously constructed. This library integrates industry-wide design standards, engineering development constraints, system architecture topology compliance rules, and element association matching criteria to form a standardized set of constraint rules. These constraints can be categorized into two types: hard mandatory constraints and soft recommended constraints. Hard mandatory constraints are rigid rules that the architecture design must meet, including topology specifications, parameter threshold boundaries, and legal association requirements. Soft recommended constraints are reference rules based on industry-wide best practices and historical engineering adaptation experience. This domain-specific constraint library provides a unified benchmark for subsequent design element extraction, intelligent association matching, architecture conflict verification, and change tracing and deduction.

[0037] S2: Automatic Extraction of Design Elements and Structured Representation of Five-Tuples This step employs a joint extraction algorithm that combines rule matching and semantic parsing to automatically identify entity elements and associated attributes from requirements documents, design specifications, architecture description documents, and parameter configuration tables.

[0038] Among them, rule matching relies on domain keywords and element feature rules to match the design entities and corresponding attributes explicitly stated in the document; semantic parsing combines contextual semantic relationships to mine the element association information implicitly stated in the text; and the two together achieve the complete extraction of explicit and implicit design elements and attributes.

[0039] Based on the design element standard dictionary built by S1, synonym mapping is carried out, and then unit normalization, format regularization and duplicate element removal are performed in sequence to complete the standardization and regularization of multi-source heterogeneous design elements.

[0040] All the standardized design elements are encapsulated into a standardized five-tuple general structured format. This five-tuple is a unified expression framework that adapts to all categories of design elements, and includes, in order: unique identifier ID, element name, element type, attribute set, and element source.

[0041] Among them, the attribute set is used to contain the attribute content of the corresponding category design elements themselves. The attributes of each type of element follow the corresponding attribute template specification defined in S1. The five-tuple belongs to the outer unified encapsulation structure of all categories of design elements. It belongs to a different level definition from the exclusive inner attribute template of each type of element. The two are not equivalent to each other.

[0042] By using the above-mentioned unified structured encapsulation method of five tuples, multi-source heterogeneous design elements with scattered sources and different expression forms are transformed into a unified and regular structured data form, realizing machine-analyzable and computable semantic expression of multi-source heterogeneous design elements.

[0043] The following is a typical example of a structured representation of a quintuple: (P001, reconnaissance platform, system elements, {endurance: 120min, maximum range: 50km}, engineering system design specifications). (C002, Data Link Communication Component, Component Elements, {Rated Bandwidth: 20Mbps, Maximum Transmission Latency: ≤100ms}, General Aviation Interface Standard). (F003, Autonomous Target Detection, Functional Elements, {Minimum Trigger Detection Distance: ≥1km}, System Functional Design Specifications); (I004, Probe data interaction interface, interface elements, {transmission message format: standard structured message}, cross-platform interface specification). (R005, Communication bandwidth constraint, constraint element, {System rated bandwidth limit: 20Mbps}, Domain specification constraint library). (E006, Regional detection coverage, performance elements, {Design target value: ≥85%}, System performance index specifications).

[0044] After the five-tuple of elements is encapsulated, a preliminary validity check is performed, mainly verifying the completeness of the five-tuple fields, the standardization of the format, and the uniqueness of the elements. Invalid element entries with missing fields, disordered formats, and duplicate redundancies are removed to form a regular and unified standardized set of design elements.

[0045] S3: Building a Knowledge Graph for System Architecture Design Using the seven categories of design elements standardized by S1 and S2 as independent graph nodes and the internal logical connections of the system architecture as directed relation edges, a complete system architecture design knowledge graph is constructed using a graph database.

[0046] The system pre-defines six core relational categories: composition, dependency, constraint, performance support, interaction, and evolutionary influence. Each graph node is configured with inherent attributes, data types, and value ranges conforming to the S1 attribute template specification. These inherent attributes correspond to the basic features, component features, interface features, and unique business attribute fields defined by various design elements in S1. Weight coefficients, initial confidence levels, and activation conditions are configured for each directed relation edge. All relevant data is uniformly stored in the graph database, supporting subsequent rapid path traversal, relational queries, and logical deduction. The knowledge graph supports real-time incremental updates and dynamic modifications.

[0047] System architecture design knowledge graph hierarchical structure, such as Figure 2As shown, the overall architecture adopts a three-layer progressive structure: the top layer is the system architecture layer, corresponding to the overall architecture view of the complex system; the middle layer is the element node layer, containing seven categories of design element nodes: system, component, interface, function, constraint, performance, and interaction; the bottom layer is the relationship and constraint layer, which sets up six core relationship sets: composition, dependency, constraint, performance support, interaction, and evolutionary impact, as well as a domain specification constraint library. The knowledge graph uses various types of design elements as entity nodes and the relationships between elements as directed edges; the element entities and their relationships all comply with the requirements of the domain specification constraint library, completing the compliant modeling of the overall architecture.

[0048] The six core relationships constructed in this embodiment are explained as follows: Compositional relationships: Representing the hierarchical and subordinate relationships between the whole and its parts. For example, a reconnaissance platform includes environmental perception components, data link communication components, and mission control components. Dependency: Represents the dependent relationship of the implementation of a function. That is, if the dependent object is missing, the corresponding function cannot be implemented normally. For example, the autonomous target detection function depends on the environmental perception component and the detection data interaction interface. Constraint relationships: These represent the parameter-limited associations between constraint elements and entity elements and functional elements. For example, data link communication components are subject to both communication bandwidth constraints and transmission delay constraints, and attribute parameters must meet constraint threshold requirements. Performance support relationship: Characterizes the support relationship between functional elements and performance indicators. Functional performance directly determines the achievement of the corresponding performance indicator. For example, multi-machine collaborative positioning function supports the improvement of regional detection coverage. Interaction relationship: Represents the bidirectional / one-way transmission and exchange of data and instructions between elements, such as the reconnaissance platform reporting detection data to the command and control platform level by level; Evolutionary impact relationships: Characterize the transmission correlation of changes in elements, used for tracing the chain of subsequent design changes. For example, changes in the parameters of environmental perception components directly affect the autonomous detection function of the target and the performance indicators of the global detection.

[0049] Through the standardized configuration of nodes, relationships, attributes, weights, and confidence levels, a system semantic network carrier that can be traversed across the entire domain, calculated numerically, and logically deduced is constructed, providing a basic carrier for subsequent element association matching, conflict detection, and change tracing.

[0050] S4: Intelligent Association and Matching of Design Elements This step constructs a weighted attribute similarity calculation model adapted to complex system architecture design scenarios. It performs element association operations based on the aforementioned knowledge graph node attributes and domain-specific constraints. The specific weighted attribute similarity calculation formula is as follows: Sim(e1,e2) = α·Sim_type + β·Sim_attr + γ·Sim_rule In the formula: Sim_type represents the feature type matching degree, with a value of 1 for consistent feature types and a value of 0 for inconsistent feature types; Sim_attr is the Jaccard similarity of the feature attribute set, which is calculated by the ratio of the intersection to the union of the attribute sets, and the value range is [0,1]. Sim_rule is the domain specification rule matching score, which is scored based on the degree to which the element conforms to the domain constraint rules, and the value range is [0,1]. The weighting coefficients are α=0.3, β=0.4, γ=0.3, and the sum of the weighting coefficients is 1.

[0051] Based on the similarity calculation results above, historical association frequency parameters and constraint compliance parameters are introduced to correct the calculation results, resulting in the element association confidence score. The calculation formula is as follows: Conf(e1,e2) = Sim(e1,e2) ×Hist_score×Const_score In the formula: Hist_score is the historical design association frequency score, with a value range of 0.8 to 1; Const_score is the score for the element constraint rule satisfaction. A value of 1 is assigned to fully satisfied constraints, a value of 0 to 1 is assigned to partially satisfied constraints, and a value of 0 is assigned to unsatisfied constraints.

[0052] Set the threshold for tiered confidence level judgment: if the association confidence level is ≥0.85, a stable association of elements is automatically established; if the association confidence level is in the range of 0.6~0.85, candidate associations are generated and pushed; if the association confidence level is <0.6, it is judged as an invalid association and directly filtered.

[0053] The quantization matching calculation results of this embodiment are as follows: The reconnaissance platform and data link communication component: the calculated association confidence level is 0.94, reaching the high confidence threshold, and a fixed association is automatically established; Communication relay platform and command and control platform: The calculated association confidence level is 0.89, which reaches the high confidence threshold, and a fixed association is automatically established; Environmental perception component and target detection function: The calculated association confidence level is 0.78, which is in the middle confidence interval, and it is listed as a candidate association; Energy security components and task instruction execution: The calculated association confidence level is 0.42, which is lower than the minimum confidence threshold, and is therefore determined to be an invalid association.

[0054] The finalized element relationships are updated in real time to the system architecture design knowledge graph. Designers can confirm and adjust candidate relationships, quickly complete the system architecture topology construction, reduce the workload of manual sorting, and ensure the accuracy and rationality of element relationships.

[0055] S5: Automatic Detection and Consistency Verification of Architectural Association Conflicts This step employs a depth-first traversal algorithm combined with a domain rule engine to perform full path traversal and rule comparison of all nodes and all relation edges in the knowledge graph. Starting with the top-level element of the system as the traversal starting node, it covers all sub-elements and relationships layer by layer, and automatically identifies and locates five types of architectural defects: circular dependency conflict, missing dependency conflict, constraint violation conflict, performance mismatch conflict, and topology violation conflict.

[0056] This embodiment of the automated detection output includes typical conflict items, judgment criteria, and correction suggestions: 1. Missing dependency conflict: The target autonomous detection function is not bound to the necessary environmental awareness components, which does not comply with the functional dependency rules. It is recommended to add matching high-confidence related components and re-establish compliant dependencies. 2. Limit Exceeding Constraints: The architecture design requires 25Mbps of link communication bandwidth, which exceeds the specification limit of 20Mbps and does not comply with the bandwidth threshold constraint rules. It is recommended to reduce the service bandwidth load or adapt and upgrade the communication components to make the parameters meet the constraint threshold requirements. 3. Performance mismatch conflict: The current architecture has a global detection coverage of only 62%, which does not meet the design target of 85% and does not comply with the performance indicator constraints. It is recommended to optimize the existing platform and component layout configuration and add perception-related links to improve the performance indicator achievement level.

[0057] This traversal and inspection did not find any circular dependency conflicts or topology violations. The system automatically marks the conflict location, conflict type, and judgment criteria, and outputs corresponding correction suggestions based on the domain specification constraint library. After manual rectification and confirmation, a second full-domain review can be triggered, iteratively verifying until the entire domain architecture meets the domain specification constraint requirements.

[0058] S6: End-to-end traceability and simulation of the impact of design changes This step employs a forward propagation and backward tracing algorithm for relationship links. When the parameters, functional configurations, or relationships of design elements within the system architecture are added, modified, or deleted, the algorithm traverses the entire path along the architecture dependency links, relying on the six types of relationships in the knowledge graph.

[0059] The forward propagation mechanism starts with the element where the change occurred and extends the propagation range layer by layer along the relationship chain. The reverse tracing mechanism starts with the affected element and traces back to locate the source of the change, completing a full traversal based on bidirectional links. During the traversal, the mechanism automatically distinguishes between directly affected elements, indirectly affected elements, and disturbed performance indicators. Combining the correlation hierarchy with the quantified evaluation architecture topology and the overall performance fluctuation range, it generates a visualized change impact analysis report containing change information, propagation path, impact level, performance fluctuation data, and optimization suggestions.

[0060] This embodiment adjusts the upper limit of the bandwidth constraint of the data link communication component from 20Mbps to 50Mbps, triggering a global change impact simulation: Directly affected levels: Data link communication components and corresponding constraint rules are updated synchronously, and the associated architecture topology remains stable; Indirect impact level: Changes spread along interaction and dependency links to service transmission links, affecting core functions such as autonomous target detection and multi-machine collaborative positioning; Performance impact level: The connectivity rate of the communication network increased from 82% to 94%, the regional detection coverage and the overall mission success rate were optimized simultaneously, and the overall performance of the architecture met the design constraints.

[0061] After the change analysis is completed, the updated element parameters, relationships and performance evaluation results will be incrementally updated to the system architecture design knowledge graph in sync, so as to realize full-link traceability and quantitative evaluation of complex system architecture design changes and form closed-loop iterative management.

[0062] The above description is merely a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. It can be used in various other combinations, modifications, and environments, and can be altered within the scope of the concept described herein through the above teachings or related technologies or knowledge. Modifications and variations made by those skilled in the art that do not depart from the spirit and scope of the present invention should be within the protection scope of the appended claims.

[0063] It should be noted that, for the sake of simplicity, the foregoing method embodiments are described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

Claims

1. A method for element association verification and change traceability in complex system architecture design, characterized in that, include: The standardization and organization of multiple design elements forms a standard dictionary of design elements, integrates various domain architecture constraint specifications to build a domain specification constraint library, and extracts design elements from multiple source documents and encapsulates them into a five-tuple format; the standardization and organization includes classification and sorting, terminology normalization, and unified definition of attribute templates. A knowledge graph for system architecture design is constructed using five-tuple design elements as nodes and architectural logical relationships as directed edges. Multiple types of core relationships are defined and attributes and weights are configured. Based on the standard dictionary of the design elements and the domain specification constraint library, a multi-dimensional weighted similarity model is used to complete element association matching and confidence determination. A rule engine combining a traversal algorithm with the domain specification constraint library is used to compare the entire path of the knowledge graph to identify and locate multiple types of architectural conflicts; when design elements change, the changes propagate forward and backward along the core relationships to distinguish the impact level and quantitatively evaluate the fluctuation of architectural performance.

2. The element association verification and change traceability method for complex system architecture design according to claim 1, characterized in that, The standardization and regularization of multiple design elements includes: For various design elements, synonym sorting and unified mapping are performed to determine a unique standard name for each design element; standardized attribute templates are developed for different categories of design elements, clarifying the fixed attribute fields and data format specifications for each type of element, and design elements of the same category are all structured using the corresponding unified attribute template; By combining rule matching and semantic parsing, entity elements and associated attributes are identified from multi-source design documents. Rule matching relies on domain keywords and element feature rules to match the explicitly stated design entities and their corresponding attributes, while semantic parsing combines contextual semantic association to mine the implicit element association information in the text. The extracted elements undergo preliminary validity checks, verifying field completeness, format standardization, and element uniqueness, and eliminating invalid element entries with missing fields, incorrect formats, or duplicates.

3. The method for element association verification and change traceability for complex system architecture design according to claim 1, characterized in that, The quintuple format includes: The unique identifier field is used to assign a globally unique identification code to each design element; The feature name field is used to record the standard naming of features after terminology normalization; The feature type field is used to identify the specific category to which a feature belongs. The attribute collection field is used to store the various attribute contents corresponding to the features. The attributes of each feature itself follow the attribute template specifications of the corresponding category. The Feature Source field records the original document or specification source from which the feature originated.

4. The method for element association verification and change traceability for complex system architecture design according to claim 1, characterized in that, The domain specification constraint library includes: Hard constraints are rigid rules that the architecture design must meet. These hard constraints include topology specifications, parameter threshold boundaries, and legal association requirements. Soft recommendation constraints serve as reference rules for industry-wide best practices and historical engineering adaptation experience.

5. The element association verification and change traceability method for complex system architecture design according to claim 1, characterized in that, The various types of core relationships include: Compositional relationships represent the hierarchical and subordinate relationships between the whole and its parts; Dependency relationships represent the dependent associations that enable the implementation of functions; that is, functions cannot be implemented normally if the corresponding objects are missing. Constraint relationships characterize the parameter-limited associations between constraint elements and entity elements and functional elements; The performance support relationship characterizes the supporting relationship between functional elements and performance indicators. Functional performance directly determines the effectiveness of the corresponding performance indicators. Interaction relationships represent the bidirectional or unidirectional transmission and exchange of data and instructions between elements; Evolutionary influence relationships characterize the transmission correlation of element changes and are used for tracing design change links.

6. The element association verification and change traceability method for complex system architecture design according to claim 1, characterized in that, The multi-dimensional weighted similarity model includes: The feature type matching degree calculation module is used to determine whether the types of the features to be matched are consistent; The attribute set similarity calculation module is used to calculate the ratio of the intersection to the union of attribute sets; The domain specification rule matching degree calculation module is used to calculate the degree to which elements conform to domain constraint rules.

7. The method for element association verification and change traceability for complex system architecture design according to claim 1, characterized in that, The completion of element association matching and confidence determination includes: Based on the similarity calculation results, historical association frequency parameters and constraint compliance parameters are introduced to correct the calculation results and obtain the element association confidence. A multi-level confidence threshold is set to divide the confidence interval into high-confidence, medium-confidence, and low-confidence intervals; the matching results in the high-confidence interval automatically establish stable element associations; the matching results in the medium-confidence interval generate candidate associations and push them; the matching results in the low-confidence interval are judged as invalid associations and directly filtered. The finalized relationships between elements are updated in real time to the system architecture design knowledge graph.

8. The method for element association verification and change traceability for complex system architecture design according to claim 1, characterized in that, The identification and location of multiple types of architectural conflicts includes: Starting with the top-level elements of the system as the traversal starting node, it covers all sub-elements and their relationships layer by layer; it automatically identifies and locates five types of architectural defects: circular dependency conflicts, missing dependency conflicts, constraint violation conflicts, performance mismatch conflicts, and topology violation conflicts. Automatically label conflict locations, conflict types, and judgment criteria, and output corresponding correction suggestions based on the domain specification constraint library; after the rectification is completed and confirmed, trigger a second full-domain review, iteratively verifying until the entire domain architecture meets the domain specification constraint requirements.

9. The method for element association verification and change traceability for complex system architecture design according to claim 1, characterized in that, The forward propagation and reverse tracing along the core relationship include: Forward propagation starts with the changed design element as the starting node and extends the change propagation range layer by layer along the relationship link; backward tracing starts with the affected element as the base point and traces back to locate the source of the change; the entire domain is traversed through the bidirectional link of forward propagation and backward tracing. The traversal process relies on all the relationship types of the knowledge graph and performs a full path traversal along the architectural dependency links; when the parameters, functional configurations, or relationships of the design elements within the system architecture are added, modified, or deleted, the tracing process is triggered.

10. The method for element association verification and change traceability for complex system architecture design according to claim 1, characterized in that, The distinction of impact levels and the quantification of architectural performance fluctuations include: The factors affected by the change are divided into three levels: directly affected factors, indirectly affected factors, and disturbed performance indicators. The directly affected factors are those directly related to the changed factor. The indirectly affected factors are those affected after being transmitted through multiple layers of correlation. The disturbed performance indicators are performance indicators that fluctuate due to the change of factors. By combining the correlation hierarchy and coupling to quantify and evaluate the fluctuation range of the architecture topology and overall performance, a visualized change impact analysis report is generated. The report includes change information, transmission path, impact hierarchy, performance fluctuation data, and optimization suggestions. The updated element parameters, relationships, and performance evaluation results are synchronously and incrementally updated to the system architecture design knowledge graph, forming a closed-loop iterative management and control.

Images