SysML architecture model based superset model generation method, device and equipment and medium

Abstract

This invention discloses a method, apparatus, device, and medium for generating superset models based on SysML architecture models. The method includes: generating candidate architecture models that meet the similarity requirements of each model variant; completing the missing hierarchical connections between elements in the candidate architecture models to obtain a full architecture model containing all model elements and all element relationships; dividing the model variants into multiple partitions by grouping model elements with interdependent relationships into the same partition, wherein interdependence indicates that at least two model elements exist simultaneously in at least two variants; identifying target partitions with substitutable functions from each partition; and adding mutation points matching the attribute characteristics of each model element in the target partition to each model element in the full architecture model, based on the attribute characteristics of each model element in the target partition, to obtain a superset model. By adopting the above technical solution, the generation efficiency of supermodels is effectively improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of systems engineering technology, and more specifically, to a method, apparatus, device, and medium for generating superset models based on the SysML architecture model. Background Technology

[0002] With the continuous evolution of system development tools and methodologies, and the increasing demands of customers, the number of variants of complex products is growing rapidly. Taking the automotive industry as an example, a large automotive company needs to manage tens of thousands of variants on a single production line. This makes variant modeling and reuse methods based on "clone-and-own" approaches unsuitable. Against this backdrop, the concept of Product Line Engineering (PLE) from the software engineering field has gradually been extended to the system engineering field and combined with Model Based Systems Engineering (MBSE) to form Model-based Product Line Engineering (MBPLE). In industrial practice, the same system often has multiple configurations or versions, corresponding to multiple interrelated but distinct model variants. A superimposed model is an architectural model that can comprehensively encompass the commonalities and variability information of all variants, and it has significant value in product line reuse and configuration management.

[0003] However, generating superset models from large-scale variants usually requires a significant investment of human resources, and in the MBPLE field, there is still a significant gap in the automated generation of superset models. On the other hand, in the Software Product Line (SPL) field, some methods for automatically generating software product line architectures have been proposed, but they still have the following limitations: (1) most of them use source code as a variant expression, and few studies focus on model-based architecture generation; (2) the generated software architecture only contains model elements without introducing mutation points, which means that the generated architecture model cannot directly support product line configuration management.

[0004] Therefore, there is an urgent need for a highly automated solution that can be seamlessly integrated with the MBSE toolchain and produces a fully reusable superset model. Summary of the Invention

[0005] This invention provides a method, apparatus, device, and medium for generating superset models based on the SysML architecture model, which effectively improves the generation efficiency of supermodels.

[0006] In a first aspect, this invention provides a method for generating superset models based on the SysML architecture model, the method comprising: Obtain a set of model variants based on the SysML architecture, and generate a candidate architecture model whose similarity to each model variant in the set of model variants meets the requirements. Complete the missing hierarchical connections between elements in the candidate architecture model to obtain a full architecture model that includes all model elements and all element relationships. Model elements with interdependent relationships in the model variant set are grouped into the same partition to divide each model variant into multiple partitions, wherein each partition includes at least one model element, and the interdependence means that at least two model elements exist simultaneously in at least two variants; Identify target partitions with alternative functions from among the partitions; Based on the attribute characteristics of each model element in the target partition, mutation points matching its attribute characteristics are added to each model element in the target partition of the full architecture model to obtain a superset model.

[0007] Optionally, generating candidate architecture models that meet the similarity requirements with each model variant in the model variant set includes: The set of model variants based on the SysML architecture is used as the initial population, and the similarity is used as the fitness value to perform genetic operations based on a genetic algorithm. The offspring population generated by the genetic operation is merged with its corresponding parent population to form the initial population for the next iteration. The genetic operation based on the genetic algorithm is then performed again until a candidate architecture model is generated that meets the similarity requirement with each model variant in the model variant set. The genetic operation includes population initialization, selection, crossover, and / or mutation operations.

[0008] Optionally, the model similarity is calculated using the following formula:

[0009] in, MS Indicates model similarity, Variants It is the input model variant group. v It is a model variant. diff This represents the difference between two variant models. architecture_model This represents the candidate architecture model.

[0010] Optionally, the missing hierarchical connections between elements in the candidate architecture model are filled in, including: An ontology model knowledge base is established based on the Web Ontology Language (OWL). The ontology model knowledge base includes instance data and semantic relationships of each model variant. The semantic relationships include hierarchical relationships between instances and data attributes, and the data attributes include parent type attributes and child type attributes. Based on the instance data and semantic relationships of each model variant in the ontology model knowledge base, the missing generalization relationships between elements in the candidate architecture model are completed.

[0011] Optionally, model elements in the model variant set that have mutual dependencies are grouped into the same partition, including: Each model variant in the model variant set is decoded into a set of primitives, each primitive representing a model element; The model variants in the model variant set are traversed sequentially. If at least two model elements in the currently traversed model variant exist simultaneously in any other model variant, then the at least two model elements are determined to be model elements with mutual dependency. Model elements in the model variant set that have interdependent relationships are grouped into the same partition.

[0012] Optionally, identify functionally alternative target partitions from among the partitions, including: For each model variant in the model variant set, if there exist at least two partitions whose intersection is an empty set and whose union is the model variant set, then the at least two partitions are identified as target partitions with a substitutable relationship.

[0013] Optionally, based on the attribute characteristics of each model element in the target partition, mutation points matching its attribute characteristics are added to each model element in the full architecture model of the target partition, including: Iterate through each element in the target partition in sequence. If the attribute feature of the currently traversed model element is an initial attribute, then the mutation point type added to the currently traversed model element is an initial attribute mutation point. The initial attribute includes any one of string, number, or boolean type. If the attribute feature of the currently traversed model element is not the initial attribute, and if the attribute feature of the currently traversed model element is another element, then the mutation point type added to the currently traversed model element is an element attribute mutation point. If the attribute characteristics of the currently traversed model element are not those of other elements, and if the currently traversed model element belongs to a variable partition, then the mutation type added to the currently traversed model element is "existence of mutation point". Here, the variable partition is any partition other than the common partition, and the common partition is the partition that belongs to all variants.

[0014] Secondly, embodiments of the present invention also provide a superset model generation apparatus based on the SysML architecture model, comprising: The candidate architecture model generation module is used to obtain a set of model variants based on SysML architecture and generate candidate architecture models whose similarity to each model variant in the set of model variants meets the requirements. The full architecture model generation module is used to fill in the missing hierarchical connection relationships between elements in the candidate architecture model to obtain a full architecture model containing all model elements and all element relationships. The partitioning module is used to divide model elements with mutual dependencies in the model variant set into the same partition, so as to divide each model variant into multiple partitions, wherein each partition includes at least one model element, and the mutual dependency means that at least two model elements exist simultaneously in at least two variants; The target partition identification module is used to identify target partitions with replaceable functions from each partition; The superset model generation module is used to add mutation points that match the attribute characteristics of each model element in the target partition to each model element in the full architecture model, so as to obtain a superset model.

[0015] Optional, a candidate architecture model generation module, specifically used for: The set of model variants based on the SysML architecture is used as the initial population, and the similarity is used as the fitness value to perform genetic operations based on a genetic algorithm. The offspring population generated by the genetic operation is merged with its corresponding parent population to form the initial population for the next iteration. The genetic operation based on the genetic algorithm is then performed again until a candidate architecture model is generated that meets the similarity requirement with each model variant in the model variant set. The genetic operation includes population initialization, selection, crossover, and / or mutation operations.

[0016] Optionally, the model similarity is calculated using the following formula:

[0017] in, MS Indicates model similarity, Variants It is the input model variant group. v It is a model variant. diff This represents the difference between two variant models. architecture_model This represents the candidate architecture model.

[0018] Optional, full architecture model generation module, specifically used for: An ontology model knowledge base is established based on the Web Ontology Language (OWL). The ontology model knowledge base includes instance data and semantic relationships of each model variant. The semantic relationships include hierarchical relationships between instances and data attributes, and the data attributes include parent type attributes and child type attributes. Based on the instance data and semantic relationships of each model variant in the ontology model knowledge base, the missing generalization relationships between elements in the candidate architecture model are completed.

[0019] Optional, partitioning module, specifically used for: Each model variant in the model variant set is decoded into a set of primitives, each primitive representing a model element; The model variants in the model variant set are traversed sequentially. If at least two model elements in the currently traversed model variant exist simultaneously in any other model variant, then the at least two model elements are determined to be model elements with mutual dependency. Model elements in the model variant set that have interdependent relationships are grouped into the same partition.

[0020] Optional, the target partition identification module is specifically used for: For each model variant in the model variant set, if there exist at least two partitions whose intersection is an empty set and whose union is the model variant set, then the at least two partitions are identified as target partitions with a substitutable relationship.

[0021] Optional, superset model generation module, specifically used for: Iterate through each element in the target partition in sequence. If the attribute feature of the currently traversed model element is an initial attribute, then the mutation point type added to the currently traversed model element is an initial attribute mutation point. The initial attribute includes any one of string, number, or boolean type. If the attribute feature of the currently traversed model element is not the initial attribute, and if the attribute feature of the currently traversed model element is another element, then the mutation point type added to the currently traversed model element is an element attribute mutation point. If the attribute characteristics of the currently traversed model element are not those of other elements, and if the currently traversed model element belongs to a variable partition, then the mutation type added to the currently traversed model element is "existence of mutation point". Here, the variable partition is any partition other than the common partition, and the common partition is the partition that belongs to all variants.

[0022] Thirdly, embodiments of the present invention also provide a computing device, comprising: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the superset model generation method based on the SysML architecture model provided in any embodiment of the present invention.

[0023] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the superset model generation method based on the SysML architecture model provided in any embodiment of the present invention.

[0024] The technical solution provided in this invention, after merging input model variants to generate a candidate architecture model, can optimize the candidate architecture model to generate a full architecture model containing all elements and the relationships between elements. Then, a partitioning algorithm is used to analyze the partitions between model variants and identify target partitions within each partition whose functional attributes have substitutable relationships. Mutation points are added to the elements of the target partitions in the full architecture model, forming a complete and reusable superset model. By adopting the above-described bottom-up automatic extraction of superset models from SysML-based model variants, the superset model generation process is highly automated, significantly reducing the time and manpower costs required to build the superset model and improving overall development efficiency. Furthermore, using SysML model variants and superset models as input and output respectively, it can be implemented as a plugin, compatible with mainstream modeling tools such as Cameo Systems Modeler (a cross-platform collaborative model-based systems engineering environment), and seamlessly integrated with existing MBSE processes, facilitating its widespread application in practical engineering. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1a A flowchart of a superset model generation method based on SysML architecture model provided in Embodiment 1 of the present invention; Figure 1b This is a schematic diagram comparing the architecture model before and after element completion provided in Embodiment 1 of the present invention; Figure 1c This is a schematic diagram of the main body provided in Embodiment 1 of the present invention; Figure 1d This is an example diagram of the partition recognition algorithm provided in Embodiment 1 of the present invention; Figure 1e Formal context diagram; Figure 1f This is a schematic diagram of a superset model of a car provided in Embodiment 1 of the present invention; Figure 2 This is a structural block diagram of the superset model generation device based on the SysML architecture model provided in Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the structure of a computing device provided in Embodiment 3 of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] It should be noted that the terms "comprising" and "having," and any variations thereof, in the embodiments and drawings of this invention are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the steps or units listed, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices.

[0029] This invention discloses a method, apparatus, device, and medium for generating superset models based on the SysML architecture model. These will be described in detail below.

[0030] Example 1 Figure 1a This is a flowchart of a superset model generation method based on the SysML architecture model provided in Embodiment 1 of the present invention. This method can be applied to the construction of supermodels for various industrial products (e.g., automobiles, aircraft, etc.). The method provided in this embodiment can be executed by a superset model generation device based on the SysML architecture model, which can be implemented through software and / or hardware. Figure 1a As shown, the method provided in this embodiment specifically includes: S110. Obtain a set of model variants based on the SysML architecture, and generate a candidate architecture model whose similarity to each model variant in the set meets the requirements.

[0031] Among them, SysML (Systems Modeling Language) architecture refers to the system architecture model built using the SysML language. For example, the SysML architecture of an automotive system takes the car as the target system and uses the SysML language specification to describe the overall structure of the car, the behavior of each component, functional requirements and constraints. The core consists of "blocks (corresponding to each component of the car)" and "connectors (corresponding to the connection / interaction relationship between components of the car)". It can intuitively and standardly present the relationship between automotive components such as the engine, transmission, and chassis, as well as the overall operating logic of the car. It is widely used in the design, analysis and verification of automotive systems.

[0032] A set of model variants based on the SysML architecture is a collection of SysML architecture models that have differentiated characteristics but inherit the core logic of the base model, generated by controllable adjustments to local structures, parameters, or connections (such as adding or deleting components, modifying types, and adjusting connection methods) based on a pre-defined SysML base architecture model. For example, based on a pre-defined automotive SysML base architecture model, a set of SysML architecture models with differentiated characteristics but inheriting the core logic of the base automotive architecture can be generated by controllable adjustments to automotive components (blocks), component connections (connectors), or parameters (such as adding or deleting reversing radars, changing engine types, and adjusting the connection method between the transmission and the engine).

[0033] For example, generating a candidate architecture model that meets the required similarity to each model variant in the model variant set can be achieved as follows: The model variant set is analyzed to extract core features common to all variants, such as common elements and common connections. Then, the common features are fused to construct the initial architecture of the candidate model (e.g., integrating the core components and basic connections common to all variants to form an initial general vehicle architecture). The initial candidate model is iteratively optimized. During the iteration process, if the similarity between the initial candidate model and the model variants is less than a preset similarity threshold, the initial candidate model is adjusted, such as by adding elements and / or connections (e.g., adjusting the engine model, optimizing the power connection method, etc.), to ensure that the similarity with all model variants after adjustment reaches the preset similarity threshold, thus obtaining the final candidate architecture model.

[0034] For example, generating candidate architecture models that meet the required similarity to each model variant in the model variant set can also be achieved in the following way: The initial population is a set of SysML-based model variants, and the similarity score is used as the fitness value. Genetic operations based on a genetic algorithm are performed. The offspring generated by these operations are merged with their corresponding parent populations to form the initial population for the next iteration. This process continues until candidate architecture models that achieve the required similarity to all model variants in the set are generated. The initial population can be constructed by copying each of the input SysML-based model variants. Each variant is included at least once, and repeated copying of variants is allowed until a predetermined population size is reached.

[0035] Specifically, model similarity is calculated using the following formula:

[0036] in, MS Indicates model similarity, Variants It is the input model variant group. v It is a model variant. diff This represents the difference between two variant models. architecture_model This represents the candidate architecture model.

[0037] In this embodiment, the genetic operations based on the genetic algorithm include selection, crossover, and / or mutation operations.

[0038] Specifically, the cross-operation starts with two model variants and generates two sub-models. One sub-model merges the differences between candidate models to generate a model containing more elements, gradually evolving towards a system architecture model; the other sub-model eliminates the differences between parent models, aiming to resolve potential element conflicts between candidate architectures. The generated sub-models are then incorporated into the group, serving as the basis for subsequent operations.

[0039] The mutation operation aims to modify each parent model once. Starting with two intermediate candidate architecture models generated during the iteration process, this operation generates two sub-models. Specifically, if there are differences between model X and model Y, one sub-model will randomly integrate an element unique to model Y into model X; the other sub-model will be based on model Y and randomly integrate an element unique to model X.

[0040] By employing a genetic algorithm to construct candidate architecture models, compared to traditional iterative optimization methods, this approach efficiently searches for candidate models that satisfy a preset threshold constraint of similarity to all variants in the model variant set through a systematic process of encoding, population initialization, genetic operations (selection, crossover, mutation), and iteration termination. This ensures that the generated candidate architecture models not only take into account the core architectural features of all variants (such as core components and power connection relationships in automobiles like engines and transmissions), but also guarantee the rationality, simplicity, and engineering practicality of the architecture. Furthermore, the algorithm can flexibly adapt to the diverse combinations and connection adjustments of variant components without requiring repeated manual fine-tuning, significantly improving the efficiency of candidate model generation, reducing the workload and cost of architecture design, and providing efficient and reliable technical support for subsequent superset model extraction.

[0041] S120. Complete the missing hierarchical connections between elements in the candidate architecture model to obtain a full architecture model containing all model elements and all element relationships.

[0042] Those skilled in the art will understand that merging architectural models using a genetic algorithm can yield a candidate architectural model most similar to the input variants, but this model may suffer from missing structural relationships. Taking a car system as an example, in different model variants, "engine" may correspond to different types such as "gasoline-powered engine," "diesel-powered engine," and "hybrid engine." In an ideal superset model, these types should be subtypes of "engine" and organized through generalization relationships. However, as... Figure 1b As shown, in the search-generated architecture model, usually only one subtype is retained, and the remaining subtypes exist as isolated modules, lacking structural association with the parent type. Therefore, it is necessary to fill in the missing hierarchical connection relationship between elements in the candidate architecture model.

[0043] Optionally, an instance-based reasoning-based relation completion method can be used to complete the aforementioned relations in the candidate architecture model. This method constructs an ontology knowledge base and utilizes the type mapping relationships in the existing system to reason and complete the missing generalization relations.

[0044] Specifically, an ontology model knowledge base can be built based on OWL (Web Ontology Language). This ontology model knowledge base includes instance data of each model variant and its semantic relationships. The semantic relationships include hierarchical relationships between instances and data attributes. The data attributes include parent type attributes and child type attributes to construct the ontology knowledge base. Figure 1cThis is a schematic diagram of an ontology. In the ontology model, instance types are used to describe the structural knowledge extracted from existing models. Each instance records a set of mapping relationships between parent types and their corresponding subtypes. This mapping is collected from existing products. Parent types represent more abstract model elements, such as "engine"; subtypes represent the specific implementation of the parent type, such as "gasoline-powered engine" and "diesel-powered engine".

[0045] Then, based on the instance data and semantic relationships of each model variant in the ontology model knowledge base, the missing generalization relationships between elements in the candidate architecture model are inferred and completed. Optionally, the instance-based mapping rules implemented using ontology reasoning can be written in SWRL (Semantic Web Rule Language) format, and the rules are as follows: An instance (c) -> has a subtype (c, s) -> has a generalization relation (s, p) For example, for each isolated subtype element in the candidate architecture model, such as a diesel engine, by searching the ontology, we can identify the parent type with which it has a connection, such as engine. Once the parent type is found, the generalization relationship between the subtype and the parent type in the candidate architecture model can be completed. For example... Figure 1b In the model, the generalization relationship between the two subtype elements, "diesel power" and "hybrid power," and their parent element (engine) is completed; the generalization relationship between the two subtype elements, "automatic" and "manual," and their parent element (transmission) is also completed. In practical application, this can be achieved by using directed edges to complete the hierarchical connection relationship between parent and child nodes.

[0046] By using the above rules, isolated subtypes in candidate architecture models can be automatically identified, and their generalization relationships with parent types can be completed, thereby generating a complete full architecture model.

[0047] S130. Divide the model elements in the model variant set that have mutual dependencies into the same partition, so as to divide each model variant into multiple partitions.

[0048] Each partition includes at least one model element, and the interdependence means that at least two model elements exist simultaneously in at least two variants.

[0049] For example, grouping model elements with interdependent relationships in a set of model variants into the same partition can include the following steps 1-3: 1. Decode each model variant in the model variant set into a set of primitives, each primitive representing a model element.

[0050] For example, the formal representation of primitives parsed in the SysML model is: [Atomic_Element_Type]Name Additional_Information Here, Atomic_Element_Type represents the element type, Name includes the SysML model name and element name, and Additional_Information represents optional additional information.

[0051] 2. Iterate through the model variants in the model variant set in turn. If at least two model elements in the currently iterated model variant exist in any other model variant, then determine the at least two model elements as model elements with mutual dependency.

[0052] 3. Group model elements with interdependent relationships in the model variant set into the same partition.

[0053] Furthermore, elements that do not have mutual dependencies are divided into different partitions, and each partition is ultimately a set of multiple elements.

[0054] Specifically, Figure 1d This is an example diagram of the partition recognition algorithm provided in Embodiment 1 of the present invention. Figure 1d As shown, given a set of variants ALLV containing four model variants, each of the four variants is represented as an ellipse, and rhombuses are the elements within these variants. A similarity measure is established between elements when there are interdependencies between elements from different variants. If the intersection between different variants contains at least two interdependent model elements, then the intersection can be considered a common partition between the two variants. For example, partition 1 is a partition of elements shared across all variants.

[0055] Furthermore, the partitions that appear in all variations are common parts, for example... Figure 1d Partition 1 in the diagram. The remaining parts besides the common parts are variable parts, for example... Figure 1d Partitions 2 through 8 in the middle.

[0056] S140. Identify target partitions from the partitions whose functions can be replaced.

[0057] As an optional implementation method, in this embodiment, the target partition can be identified in the following way: For each model variant in the model variant set, if there exist at least two partitions whose intersection is an empty set and whose union is the model variant set, then these at least two partitions are identified as target partitions with a substitutable relationship. This can be achieved through the following formula: and

[0058] in, , Let represent the i-th partition, n represent the total number of partitions, and G represent the set of model variants.

[0059] Furthermore, the aforementioned substitutability relationships can be obtained by creating a formal context graph, for example, for... Figure 1d The partition diagram shown corresponds to the formal context diagram as follows: Figure 1e As shown in Figure 1e, it can be seen that there is a substitution relationship between partitions {partition 2, partition 7} and partitions {partition 3, partition 5}.

[0060] As another alternative implementation method, the core functional attributes and corresponding functional parameters of each partition can be determined first (for example, based on the OWL ontology). Then, the functional attribute feature vectors of each partition are constructed, and the functional attribute feature vectors of each partition are compared pairwise. A functional attribute substitutability threshold is set. When the functional attribute similarity of two partitions reaches the threshold, and the core functions they perform are the same, that is, the same or similar system functions can be achieved by replacing the two partitions, and the replacement does not affect the rationality of the overall system architecture and the safety of operation, the two partitions can be determined as target partitions with substitutable functional attributes. For example, for a transmission, automatic, manual, and semi-automatic are substitutable partitions.

[0061] In this embodiment, by identifying target partitions with substitutable functional attributes, the components that may differ on the ontology can be obtained, i.e., the carrier for adding mutation points can be found. This ensures that the generated superset model can fully cover the functional requirements of each model variant and adapt to the core architectural features of all model variants. It can also flexibly accommodate the functional configuration of each substitutable partition, reducing the difficulty and workload of building the superset model.

[0062] S150. Based on the attribute characteristics of each model element in the target partition, add mutation points that match its attribute characteristics to each model element in the target partition of the full architecture model to obtain a superset model.

[0063] Among them, the mutation points refer to the specific differences in structural composition and functional configuration between each model variant and the basic architecture model or other variants in the model variant set. These differences include the replacement of the type of main body components (e.g., whether the car engine is a diesel engine, a gasoline engine, or a hybrid engine), the adjustment of connection relationships (e.g., changing the type of car power transmission flow, optimizing the interaction logic between components), and the differentiated settings of parameter configurations. These mutation points are the key identifiers for each variant to achieve unique functions and performance positioning, and are also the core basis for building the diversity of super models.

[0064] In this embodiment, the superset model refers to the full architecture model formed by adding explicit mutation points to each model variant based on the basic architecture and model variant set, and integrating the common core architecture of all model variants with the differentiated mutation points of each variant.

[0065] The attribute characteristics of model elements include initial attributes, element attributes, and attributes indicating the presence of mutation points. Initial attributes indicate that the attribute settings of model elements can be of types such as boolean, string, or number; for example, the number of doors on a car can be set to a number (2 or 4 doors). In SysML, this mutation point is represented by applying the «PrimitivePropertyVariationPoint» stereotype to the model element.

[0066] Element properties indicate that a model element's property can be set to another type of the model element, such as a part / reference property with a module type. In SysML, this variation point is represented by applying the «ElementPropertyVariationPoint» stereotype to the model element.

[0067] The ExistenceVariationPoint property indicates whether a model element may or may not exist in certain variants. In SysML, this mutation point is represented by applying the «ExistenceVariationPoint» stereotype to the model element.

[0068] Specifically, adding mutation points that match the attribute characteristics of each model element in the target partition of the full architecture model can be achieved in the following way: Iterate through each element in the target partition in turn. If the attribute of the currently iterated model element is the initial attribute, then add the mutation point type of the currently iterated model element as the initial attribute mutation point. The initial attribute can be any one of string, number or boolean type. If the attribute of the currently traversed model element is not the initial attribute, and the attribute of the currently traversed model element is another element, then the mutation point type added to the currently traversed model element is an element attribute mutation point. If the attribute characteristics of the currently traversed model element are not those of other elements, and if the currently traversed model element belongs to a variable partition, then the mutation type added to the currently traversed model element is "existence of mutation". Here, the variable partition is any partition other than the common partition, and the common partition is the partition that belongs to all variants.

[0069] Specifically, Figure 1f for Figure 1b The corresponding superset model diagram shows that when a mutation point is applied to a model element, a small (v) icon will appear on the element. The model element represented by the mutation point can be configured according to features to generate specific variant models.

[0070] In this embodiment, by adding mutation points to target partitions with substitutable functional attributes, it can be ensured that the super model can fully cover the functional requirements and structural features of all variants, avoid functional omissions, and improve the integrity and reusability of the super model.

[0071] The technical solution provided in this embodiment, after merging input model variants to generate a candidate architecture model, can optimize the candidate architecture model to generate a full architecture model containing all elements and the relationships between elements. Then, a partitioning algorithm is used to analyze the partitions between model variants and identify target partitions in each partition whose functional attributes have substitutable relationships. Mutation points are added to the elements of the target partitions in the full architecture model to form a complete and reusable superset model. By adopting the above-mentioned scheme of automatically extracting superset models from SysML-based model variants from the bottom up, the superset model generation process is highly automated, significantly reducing the time and manpower costs required to build superset models and improving overall development efficiency. Furthermore, using SysML model variants and superset models as input and output respectively, it can be implemented as a plug-in, is compatible with mainstream modeling tools such as Cameo Systems Modeler, and seamlessly integrates with existing MBSE processes, facilitating its promotion and application in practical engineering.

[0072] Example 2 Figure 2 This is a structural block diagram of a superset model generation device based on the SysML architecture model provided in Embodiment 2 of the present invention, as shown below. Figure 2 As shown, the device includes a candidate architecture model generation module 210, a full architecture model generation module 220, a partitioning module 230, a target partitioning identification module 240, and a superset model generation module 250, wherein... The candidate architecture model generation module 210 is used to obtain a set of model variants based on SysML architecture and generate a candidate architecture model whose similarity to each model variant in the set of model variants meets the requirements. The full architecture model generation module 220 is used to fill in the missing hierarchical connection relationships between elements in the candidate architecture model to obtain a full architecture model containing all model elements and all element relationships. Partitioning module 230 is used to divide model elements with mutual dependencies in the model variant set into the same partition, so as to divide each model variant into multiple partitions, wherein each partition includes at least one model element, and the mutual dependency means that at least two model elements exist simultaneously in at least two variants; The target partition identification module 240 is used to identify target partitions with alternative functions from among the partitions; The superset model generation module 250 is used to add mutation points that match the attribute characteristics of each model element in the target partition in the full architecture model to each model element in the target partition, so as to obtain a superset model.

[0073] Optionally, the candidate architecture model generation module 210 is specifically used for: The set of model variants based on the SysML architecture is used as the initial population, and the similarity is used as the fitness value to perform genetic operations based on a genetic algorithm. The offspring population generated by the genetic operation is merged with its corresponding parent population to form the initial population for the next iteration. The genetic operation based on the genetic algorithm is then performed again until a candidate architecture model is generated that meets the similarity requirement with each model variant in the model variant set. The genetic operation includes population initialization, selection, crossover, and / or mutation operations.

[0074] Optionally, the model similarity is calculated using the following formula:

[0075] in, MS Indicates model similarity, Variants It is the input model variant group. v It is a model variant. diff This represents the difference between two variant models. architecture_model This represents the candidate architecture model.

[0076] Optional, the full architecture model generation module 220, specifically used for: An ontology model knowledge base is established based on the Web Ontology Language (OWL). The ontology model knowledge base includes instance data and semantic relationships of each model variant. The semantic relationships include hierarchical relationships between instances and data attributes, and the data attributes include parent type attributes and child type attributes. Based on the instance data and semantic relationships of each model variant in the ontology model knowledge base, the missing generalization relationships between elements in the candidate architecture model are completed.

[0077] Optional, partition module 230, specifically used for: Each model variant in the model variant set is decoded into a set of primitives, each primitive representing a model element; The model variants in the model variant set are traversed sequentially. If at least two model elements in the currently traversed model variant exist simultaneously in any other model variant, then the at least two model elements are determined to be model elements with mutual dependency. Model elements in the model variant set that have interdependent relationships are grouped into the same partition.

[0078] Optional, the target partition identification module 240 is specifically used for: For each model variant in the model variant set, if there exist at least two partitions whose intersection is an empty set and whose union is the model variant set, then the at least two partitions are identified as target partitions with a substitutable relationship.

[0079] Optional, superset model generation module 250, specifically used for: Iterate through each element in the target partition in sequence. If the attribute feature of the currently traversed model element is an initial attribute, then the mutation point type added to the currently traversed model element is an initial attribute mutation point. The initial attribute includes any one of string, number, or boolean type. If the attribute feature of the currently traversed model element is not the initial attribute, and if the attribute feature of the currently traversed model element is another element, then the mutation point type added to the currently traversed model element is an element attribute mutation point. If the attribute characteristics of the currently traversed model element are not those of other elements, and if the currently traversed model element belongs to a variable partition, then the mutation type added to the currently traversed model element is "existence of mutation point". Here, the variable partition is any partition other than the common partition, and the common partition is the partition that belongs to all variants.

[0080] Example 3 Please see Figure 3 , Figure 3 This is a schematic diagram of the structure of a computing device provided in Embodiment 3 of the present invention. Figure 3 As shown, the computing device may include: Memory 701 storing executable program code; Processor 702 coupled to memory 701; The processor 702 calls the executable program code stored in the memory 701 to execute the superset model generation method based on the SysML architecture model provided in any embodiment of the present invention.

[0081] This invention discloses a computer-readable storage medium storing a computer program that causes a computer to execute the superset model generation method based on the SysML architecture model provided in any embodiment of this invention.

[0082] In various embodiments of the present invention, it should be understood that the sequence number of each process does not necessarily imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0083] In the embodiments provided by this invention, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined based on A. However, it should also be understood that determining B based on A does not mean determining B solely based on A; B can also be determined based on A and / or other information.

[0084] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0085] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-accessible memory. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several requests to cause a computer device (which can be a personal computer, server, or network device, specifically a processor in the computer device) to execute some or all of the steps of the methods described in the various embodiments of the present invention.

[0086] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, including read-only memory (ROM), random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), one-time programmable read-only memory (OTPROM), electrically-Erasable Programmable Read-Only Memory (EEPROM), compactdisc read-only memory (CD-ROM) or other optical disc storage, disk storage, magnetic tape storage, or any other computer-readable medium capable of carrying or storing data.

[0087] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of one embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing the present invention.

[0088] Those skilled in the art will understand that the modules in the apparatus of the embodiments can be distributed in the apparatus of the embodiments as described in the embodiments, or they can be located in one or more devices different from this embodiment with corresponding changes. The modules of the above embodiments can be combined into one module, or they can be further divided into multiple sub-modules.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for generating superset models based on the SysML architecture model, characterized in that, include: Obtain a set of model variants based on the SysML architecture, and generate a candidate architecture model whose similarity to each model variant in the set of model variants meets the requirements. Complete the missing hierarchical connections between elements in the candidate architecture model to obtain a full architecture model that includes all model elements and all element relationships. Model elements with interdependent relationships in the model variant set are grouped into the same partition to divide each model variant into multiple partitions, wherein each partition includes at least one model element, and the interdependence means that at least two model elements exist simultaneously in at least two variants; Identify target partitions with alternative functions from among the partitions; Based on the attribute characteristics of each model element in the target partition, mutation points matching its attribute characteristics are added to each model element in the target partition of the full architecture model to obtain a superset model.

2. The method according to claim 1, characterized in that, The generation of candidate architecture models that meet the similarity requirements with each model variant in the model variant set includes: The set of model variants based on the SysML architecture is used as the initial population, and the similarity is used as the fitness value to perform genetic operations based on a genetic algorithm. The offspring population generated by the genetic operation is merged with its corresponding parent population to form the initial population for the next iteration. The genetic operation based on the genetic algorithm is then performed again until a candidate architecture model is generated that meets the similarity requirement with each model variant in the model variant set. The genetic operation includes population initialization, selection, crossover, and / or mutation operations.

3. The method according to claim 2, characterized in that, The model similarity is calculated using the following formula: in, MS Indicates model similarity, Variants It is the input model variant group. v It is a model variant. diff This represents the difference between two variant models. architecture_model This represents the candidate architecture model.

4. The method according to claim 1, characterized in that, The step of completing the missing hierarchical connections between elements in the candidate architecture model includes: An ontology model knowledge base is established based on the Web Ontology Language (OWL). The ontology model knowledge base includes instance data and semantic relationships of each model variant. The semantic relationships include hierarchical relationships between instances and data attributes, and the data attributes include parent type attributes and child type attributes. Based on the instance data and semantic relationships of each model variant in the ontology model knowledge base, the missing generalization relationships between elements in the candidate architecture model are completed.

5. The method according to claim 1, characterized in that, The step of grouping model elements with interdependent relationships in the model variant set into the same partition includes: Each model variant in the model variant set is decoded into a set of primitives, each primitive representing a model element; The model variants in the model variant set are traversed sequentially. If at least two model elements in the currently traversed model variant exist simultaneously in any other model variant, then the at least two model elements are determined to be model elements with mutual dependency. Model elements in the model variant set that have interdependent relationships are grouped into the same partition.

6. The method according to claim 1, characterized in that, The process of identifying target partitions from each partition that can be functionally replaced includes: For each model variant in the model variant set, if there exist at least two partitions whose intersection is an empty set and whose union is the model variant set, then the at least two partitions are identified as target partitions with a substitutable relationship.

7. The method according to claim 1, characterized in that, The step of adding mutation points matching the attribute characteristics of each model element in the target partition of the full architecture model to each model element in the target partition includes: Iterate through each element in the target partition in sequence. If the attribute feature of the currently traversed model element is an initial attribute, then the mutation point type added to the currently traversed model element is an initial attribute mutation point. The initial attribute includes any one of string, number, or boolean type. If the attribute feature of the currently traversed model element is not the initial attribute, and if the attribute feature of the currently traversed model element is another element, then the mutation point type added to the currently traversed model element is an element attribute mutation point. If the attribute characteristics of the currently traversed model element are not those of other elements, and if the currently traversed model element belongs to a variable partition, then the mutation type added to the currently traversed model element is "existence of mutation point". Here, the variable partition is any partition other than the common partition, and the common partition is the partition that belongs to all variants.

8. A superset model generation device based on the SysML architecture model, characterized in that, include: The candidate architecture model generation module is used to obtain a set of model variants based on SysML architecture and generate candidate architecture models whose similarity to each model variant in the set of model variants meets the requirements. The full architecture model generation module is used to fill in the missing hierarchical connection relationships between elements in the candidate architecture model to obtain a full architecture model containing all model elements and all element relationships. The partitioning module is used to divide model elements with mutual dependencies in the model variant set into the same partition, so as to divide each model variant into multiple partitions, wherein each partition includes at least one model element, and the mutual dependency means that at least two model elements exist simultaneously in at least two variants; The target partition identification module is used to identify target partitions with replaceable functions from each partition; The superset model generation module is used to add mutation points that match the attribute characteristics of each model element in the target partition to each model element in the full architecture model, so as to obtain a superset model.

9. A computing device, characterized in that, The computing device includes: One or more processors; Storage device for storing one or more programs. When the one or more programs are executed by the one or more processors, the one or more processors implement the superset model extraction method based on the SysML architecture model as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the superset model generation method based on the SysML architecture model as described in any one of claims 1-7.

Images