Intelligent business application platform based on ontology model and engine driving

Abstract

The invention discloses an intelligent business application platform based on an ontology model and engine driving, relates to the technical field of artificial intelligence, and is characterized in that an ontology data engine constructs data services and knowledge maps with consistent semantics, eliminates data islands, and forms enterprise-level unified business information views. The ontology behavior logic engine carries out visual modeling on business rules, breaks a black box barrier, and assists agile business adjustment and AI interpretation. Multi-engine collaborative linkage is achieved, an integrated link of data, logic, functions and knowledge is broken through, and the problems of flow and data separation and low integration efficiency are solved. The knowledge generation engine realizes mining and conversion of dominant and implicit knowledge, and drives the system to continuously evolve. The intelligent interaction engine optimizes the user experience, and the system integration engine reduces the cross-system linkage cost. According to the platform, the service ontology model is used as a unified semantic core and a coordination center, the problems of fragmentation, semantic islands and the like in the prior art are effectively solved, and service full-link closed-loop intelligence is achieved.

Description

Intelligent business application platform based on ontology model and engine Technical Field

[0001] This application relates to the field of artificial intelligence technology, specifically to an intelligent business application platform based on ontology models and engine-driven architecture. Background Technology

[0002] In the process of enterprise digital transformation, related supporting technologies have formed multiple development directions, but all of these technologies have significant limitations and are unable to meet the needs of enterprises for intelligent collaboration across the entire chain. Data platforms and knowledge graphs struggle to break down semantic silos and are disconnected from business loops; low-code platforms focus on technical models, making business logic difficult for business personnel and AI to understand; process technologies such as BPM separate process rules from data models, leading to process rigidity; traditional application suites have fixed functions and inefficient integration, preventing knowledge from evolving in tandem with the system.

[0003] The enterprise digital system built upon the aforementioned technical solutions exhibits common pain points across multiple dimensions, hindering the effectiveness of transformation. At the data level, business concepts are defined in a fragmented manner, heterogeneous systems form "data silos," lacking a unified semantic view. This hinders cross-system data fusion, understanding, and correlation analysis, making it difficult to form a unified enterprise-level business information view. At the logic level, core business rules and processes are hard-coded and embedded in the program, presenting a "black box" to business personnel and creating a barrier to interpretation for AI systems, impeding agile business adjustments and intelligent evolution. At the interaction level, traditional systems rely on fixed interfaces and menus, focusing on processes rather than user tasks, failing to provide proactive and personalized services, resulting in a poor user experience. At the integration level, customized interface integration models suffer from high costs, semantic inconsistencies, and maintenance difficulties, hindering cross-system business capability orchestration and real-time linkage. At the knowledge level, business knowledge is largely confined to documents or expert knowledge, making it difficult to automatically extract tacit knowledge from operational data or systematically transform it into executable rules, resulting in a lack of self-learning and evolutionary capabilities.

[0004] Existing technologies only address single, localized pain points, lacking a unified semantic foundation and collaborative mechanisms. This fragmented approach fails to achieve closed-loop intelligence across the entire business chain. Various technologies operate independently, hindering multi-dimensional integrated modeling and collaborative evolution, leading to a predicament of localized optimization in the transformation process. There is an urgent need for a new platform architecture technology that reshapes application building methods and enables multi-dimensional integrated collaborative evolution. Summary of the Invention

[0005] To address the aforementioned technical problems, this application proposes the following technical solution: Firstly, embodiments of this application provide an intelligent business application platform based on an ontology model and engine-driven architecture, comprising: a business ontology model; an ontology data engine, an ontology function engine, an ontology behavior logic engine, a system integration engine, an intelligent interaction engine, and a knowledge generation engine that are interactively connected to the business ontology model; the business ontology model serves as the sole trusted source and information coordination hub, interacting and collaborating with the aforementioned engines through a pre-defined internal interface based on a unified semantic model; the ontology data engine provides integrated, semantically consistent data services to all engines, generating a knowledge graph of the core context; the ontology function engine provides atomic computation and service capabilities, is invoked by the ontology behavior logic engine to execute business processes, and is simultaneously orchestrated by the system integration engine to provide external services. The system provides data and business services; the ontology behavior logic engine models business processes, and its defined process models can be triggered by the intelligent interaction engine to generate user operations, or analyzed and optimized by the knowledge generation engine; the intelligent interaction engine directly provides users with a dynamic user interface driven by task requirements, uses the ontology data engine to acquire data, calls the ontology behavior logic engine to execute business operations, and enriches the interaction context based on external information accessed by the system integration engine; the knowledge generation engine continuously learns from the full-domain operation data, behavior logic execution records, and interaction feedback, and feeds the new knowledge it generates back to the business ontology model, ontology function engine, ontology behavior logic engine, and intelligent interaction engine, continuously optimizing the business ontology model and driving the continuous evolution of the system; the system integration engine ensures that the business capabilities and data of external systems can be integrated for data exchange and collaboration.

[0006] In one possible implementation, the business ontology model serves as the sole trusted source and information coordination hub. It interacts and collaborates with the aforementioned engines through a pre-defined internal interface based on a unified semantic model. This includes: the business ontology model receiving and updating business entity instance data and relationships written by the ontology data engine; the ontology function engine reading and converting the state of the business entity data written by the business ontology model; the business ontology model receiving execution fact chains from the ontology behavior logic engine, which reads the logical definitions of business operations or events from the business ontology model and writes them into execution instances; the business ontology model receiving updates to external data sources and instance data mappings from the system integration engine, which reads integration contracts from the business ontology model and converts them into the state of the business entity data; the intelligent interaction engine obtaining the business semantic context of user task understanding from the business ontology model; the business ontology model receiving newly discovered knowledge and rules from the knowledge generation engine and storing them as continuously evolving business knowledge memory for the platform; and the knowledge generation engine using the business ontology model as a deterministic knowledge source to automatically reason, generate, and write the newly discovered knowledge.

[0007] In one possible implementation, the ontology data engine provides integrated and semantically consistent data services to all engines, generating a knowledge graph of the core context, including: data modeling and graph generation based on the semantic definition of the unified business ontology model; providing raw data for the computation functions of the ontology function engine, providing context data after business semantic fusion for the logical execution of the ontology behavior logic engine, providing display data after business semantic fusion for the interface presentation of the intelligent interaction engine, and providing structured instance data after business semantic fusion for the analysis and reasoning of the knowledge generation engine; receiving external data processed by the data integration part of the system integration engine and writing it into the business ontology model and the unified data view integrated into the ontology.

[0008] In one possible implementation, before writing multi-source data into the business ontology model, the ontology data engine constructs a semantic consistency fusion objective and minimizes the fusion loss: in: The number of data sources participating in the integration, For the first The credibility weight of each data source The final attribute values ​​after fusion, i.e., the reliable results prepared for writing into the business ontology model, For the first Observations of the same business entity attribute from multiple data sources. As a measure of difference, it measures the cost of the fused value deviating from the observations of the data source. These are the weighting coefficients for semantic consistency constraints. Represents the set of constraints on the ontology. Summing each constraint instance in the table, As the main entity, and For two object entities related to the subject, For the ontology constraint set, This is a function that maps entities to graph embedding vectors. It is the square of the L2 norm.

[0009] In one possible implementation, the ontology function engine provides atomic computation and service capabilities, which are invoked by the ontology behavior logic engine to execute business processes, and simultaneously orchestrated by the system integration engine to provide data and business services to external entities. This includes: the ontology function engine defining computation functions and business service models, registering and managing them in a unified business ontology model to ensure semantic consistency; the computation functions and business services being the basic execution capability units for orchestration and invocation by the ontology behavior logic engine; when an ontology function is invoked and executed, the input or output parameter rules and business semantics of the ontology function are obtained from the business ontology model; the business services registered by the ontology function engine are discovered, orchestrated, and integrated with the business processes of external systems by the service integration component of the system integration engine; and the backend capabilities supporting the intelligent interaction engine are used to implement the specific backend functions of the behavioral logic triggered by user operations or the dynamic computational information displayed on the interface within the intelligent interaction engine.

[0010] In one possible implementation, the ontology behavior logic engine models the business process, including: defining the semantics, conditions, inputs, process logic flow, and output of business operations and business events in a visual, structured, and data-driven manner based on a unified behavior model; parsing and executing the business operation and business event model, calling the computation functions or business service capabilities of the ontology function engine, and supporting the integration and invocation of large language model services to achieve automated and intelligent execution of business processes; monitoring changes in the state of business operations, system events, or business entity data, and automatically driving the execution of the logical flow of business operations and business events when the operation or triggering conditions of the behavior logic model are met.

[0011] In one possible implementation, the system integration engine ensures that the business capabilities and data of external systems can be integrated for data exchange and collaboration, including: registering and connecting external applications, orchestrating service integration processes based on a unified business service model, realizing business process integration and real-time linkage between the platform and external systems; the platform accesses external system data sources, enabling external data and internal data to be associated, mapped and merged based on a unified business ontology model to construct a full-domain data view, and supporting the automatic reverse generation of the platform's business entities and their data models based on the external data model.

[0012] In one possible implementation, the system integration engine accesses the external data model. With business ontology model The entity-relationship mapping uses a joint optimization objective to determine the mapping function. , so that: in: For the set of nodes of the external model, For the set of edges of the external model, The set of nodes in the ontology. The set of edges in the ontology. For the optimal mapping function, This means taking the option that maximizes the objective function. , For nodes in the node set of the external model, Weights for semantic similarity terms. For semantic similarity function, external node semantic representation, This is the semantic representation of the ontology node. and For the ontology node, These are two nodes connected in the external model. For the structural consistency term weight, For the set of ontology relations, For indicator functions, For the penalty intensity weight, For mapping functions Inconsistency measure.

[0013] In one possible implementation, the intelligent interaction engine directly provides users with a task-demand-driven, dynamic user interface, including: automatically mapping business entities, attributes, relationships, and business operation metadata models to generate page components, page fields, page layouts, and operation buttons for the user interface based on the business ontology model, UI mapping framework, and page component library; parsing the user's natural language task requirements, automatically identifying relevant business entities and business operations, as well as the required page components, and selecting a component set from the page component library to dynamically assemble the task interaction interface, satisfying the coverage constraints of the task requirements and optimizing the objective function. in: This indicates finding the minimum value. Select a vector for the components, indicating which components to choose from the component library to enter the current task page. , Each dimension can only take the value 0 or 1, and there are a total of dimension, This represents the total number of components in the page component library. For the semantic requirements set of the current user task, for One of the specific requirements in the document. For demand Importance weights For component indexing, For the first Binary selection variables for each component, For the first Candidate components Representation Component Demand coverage The weighting coefficient for the interaction cost item. For components Interaction cost The weighting coefficient for redundant terms. This indicates the relationship between all different components. Summation of combinations, For the first Binary selection variables for each component, For components and The system calculates the conflict / redundancy coefficients; then, based on the requirements of the page model configuration framework and interaction engine, it dynamically assembles and presents personalized task interaction interfaces; it provides a unified page component registration model and integration framework, supporting the registration of user-developed customized pages as platform-runnable page components.

[0014] In one possible implementation, the knowledge generation engine continuously learns from the full-domain operational data, behavioral logic execution records, and interactive feedback, and feeds back the new knowledge it generates to the business ontology model, ontology function engine, ontology behavioral logic engine, and intelligent interaction engine. This continuously optimizes the business ontology model and drives the system's continuous evolution. This includes: automatically extracting structured business entity relationships, attribute constraints, business behaviors, business rules, and logical processes as known deterministic knowledge from the business ontology model and its business instance data; constructing specific prompt word templates, and based on user task scenario requirements analysis, guiding the large language model to automatically match and combine deterministic business knowledge with a built-in industrial domain professional algorithm library to perform causal reasoning, pattern discovery, and hypothesis generation, automatically discovering new business rules and implicit knowledge, and automatically generating knowledge solutions; and conducting online debugging and verification of the knowledge solutions discovered by the large language model based on the operational instance data of the business ontology or the simulation data generated by the large language model, selecting verified knowledge solutions, automatically publishing and registering them as platform-runnable knowledge service components, and automatically updating them to the business ontology model, forming a closed loop of cognitive application of the business ontology.

[0015] In this embodiment, the ontology data engine constructs semantically consistent data services and knowledge graphs, eliminating data silos and forming a unified enterprise-level business information view. The ontology behavior logic engine visualizes and models business rules, breaking down "black box" barriers and facilitating agile business adjustments and AI interpretation. Multi-engine collaboration connects data, logic, functions, and knowledge into a unified chain, resolving issues of process and data fragmentation and inefficient integration. The knowledge generation engine mines and transforms explicit and implicit knowledge, driving continuous system evolution. The intelligent interaction engine optimizes user experience, and the system integration engine reduces cross-system collaboration costs, fundamentally reshaping application construction methods and achieving multi-dimensional integrated collaborative evolution. This overcomes local optimization challenges and significantly improves the effectiveness of enterprise digital transformation. This platform uses the business ontology model as a unified semantic core and coordination hub, effectively addressing existing issues such as technology fragmentation and semantic silos, achieving closed-loop intelligence across the entire business chain. Attached Figure Description

[0016] Figure 1 is a schematic diagram of an intelligent business application platform based on an ontology model and engine-driven according to an embodiment of this application; Figure 2 is a schematic diagram of the business ontology model structure according to an embodiment of this application; Figure 3 is a flowchart of the ontology data engine according to an embodiment of this application; Figure 4 is a flowchart of the function registration mechanism and usage relationship of the ontology function engine according to an embodiment of this application; Figure 5 is a flowchart of the modeling and execution process of the ontology behavior logic engine according to an embodiment of this application; Figure 6 is a flowchart of the workflow and interaction relationship of the system integration engine according to an embodiment of this application; Figure 7 is a flowchart of the dynamic generation process of the task interface of the intelligent interaction engine according to an embodiment of this application; Figure 8 is a flowchart of the closed-loop workflow of the knowledge generation engine according to an embodiment of this application. Detailed Implementation

[0017] The present solution will now be described in conjunction with the accompanying drawings and specific embodiments.

[0018] Referring to Figure 1, the intelligent business application platform based on ontology model and engine-driven provided in this embodiment includes: a business ontology model, an ontology data engine, an ontology function engine, an ontology behavior logic engine, a system integration engine, an intelligent interaction engine, and a knowledge generation engine that interact with the business ontology model.

[0019] The business ontology model, as the platform's core technical feature, is the "digital soul" and "unified language" of the entire platform, serving as the operating system kernel for all business applications. It is a "formalized and standardized" set of definitions for all business concepts, entities, attributes, constraints, relationships, functional interfaces, behavioral logic, interaction patterns, and integration contracts within the enterprise's business domain. It provides a naturally unified business semantic environment and is the sole reliable source and overarching framework for all engine interactions, data flow, and knowledge generation.

[0020] Referring to Figure 2, this model is not a static dictionary, but a dynamic semantic association network that is parsable, reasonable, and scalable. It includes a business entity and attribute structure model, a dynamic entity relationship graph model, a computation function model, a business service model, a business operation and business event model, a behavioral logic model, UI mapping rules, and a service and data integration model. It is a "living" business model that supports dynamic computation, reasoning, and interaction, providing a "naturally unified and accurate" business semantic environment. It is the cornerstone for ensuring semantic consistency among all components within the platform, eliminating information silos, and enabling intelligent collaborative work.

[0021] It provides the semantic foundation for all engines: it is the sole reliable source and constraint for the six engines to define models, interpret data, make logical judgments, and design interactions. The ontology data engine writes business entities and relationships into it; the ontology function engine reads and transforms the state of the business entity data from it; the ontology behavior logic engine reads the logical definition of business operations or business events from it and writes it into execution instances; the system integration engine reads the integration contract from it and transforms it into the state of the business entity data; the intelligent interaction engine obtains the business semantic context of user task understanding from it; and the knowledge generation engine uses it as a deterministic knowledge source to automatically reason, generate, and write newly discovered knowledge.

[0022] Receive and persist the outputs of each engine: receive business entity instance data and relationship updates written from the ontology data engine; receive external data sources and instance data mapping updates accessed from the system integration engine; receive the execution fact chain from the ontology behavior logic engine; and receive newly discovered knowledge and rules from the knowledge generation engine, and use them as the business knowledge memory for the platform's continuous evolution.

[0023] Referring to Figure 3, the ontology data engine is responsible for abstract modeling, dynamic fusion, and unified service provision of business data based on the business ontology model. It is the platform's "data fusion center": it "translates" and "injects" heterogeneous business data into the unified business ontology model, realizes semantic fusion and dynamic association of data, and provides unified data services based on business semantics.

[0024] It provides visual modeling tools to define business entities, attributes, and relationships based on a unified semantic framework. Based on entity attribute and relationship models, it automatically generates and updates a dynamic, "business data native" knowledge graph in real time, synchronized with the business data status. It offers APIs / SDKs for "data storage, querying, and manipulation" based on business ontology semantics, shielding the underlying heterogeneous databases (relational / analytic / time-series / vector databases, etc.) and supporting complex relational queries and operations based on the business entity relationship graph.

[0025] Data modeling and graph generation are performed based on the semantic definition of the unified business ontology model. This provides raw data for the computational functions of the ontology function engine; contextual data after business semantic fusion for the logical execution of the ontology behavior logic engine; display data after business semantic fusion for the interface presentation of the intelligent interaction engine; and structured instance data after business semantic fusion for the analysis and reasoning of the knowledge generation engine. It also receives external data processed by the "Data Integration" part of the system integration engine and writes it into the business ontology model (especially updating the entity and instance data relationship graph) and integrates it into the unified ontology data view.

[0026] Before writing multi-source data into the business ontology model, the ontology data engine constructs a semantic consistency fusion objective and minimizes the fusion loss: in: The number of data sources participating in the integration, For the first The credibility weight of each data source The final attribute values ​​after fusion, i.e., the reliable results prepared for writing into the business ontology model, For the first Observations of the same business entity attribute from multiple data sources. As a measure of difference, it measures the cost of the fused value deviating from the observations of the data source. These are the weighting coefficients for semantic consistency constraints. Represents the set of constraints on the ontology. Summing each constraint instance in the table, As the main entity, and For two object entities related to the subject, For the ontology constraint set, This is a function that maps entities to graph embedding vectors. It is the square of the L2 norm.

[0027] Referring to Figure 4, the ontology function engine, as the platform's "capability component library," is based on a unified computation function model and business service model to realize the abstract modeling and encapsulation of the "observation capability" (computation function) and "change capability" (business service) of the data state of business entity instances.

[0028] It provides stateless and side-effect-free business data state reading and analysis capabilities, and supports user-defined calculation functions based on a unified calculation function model. It encapsulates the behavioral capabilities of business entities (changing but not holding the business entity instance data state), defines service technical parameters, business parameters, and mapping conversion rules between the two parameter models based on a unified business service model, thereby shielding the details of underlying technical parameters and providing a user-friendly and easily configurable interface. It also supports the integration of user-defined business services based on the unified business service model.

[0029] The computation functions and business service models defined by the ontology function engine are registered and managed in a unified business ontology model to ensure semantic consistency. When an ontology function is invoked and executed, the input / output parameter rules and business semantics of the ontology function are obtained from the business ontology model. The computation functions and business services encapsulated by the ontology function engine are the basic execution capability units for orchestration and invocation by the ontology behavior logic engine. The business services registered by the ontology function engine can be discovered, orchestrated, and integrated with the business processes of external systems by the "service integration" component of the system integration engine. The ontology function engine is the implementer of the backend specific functions of the behavioral logic triggered by user operations or the dynamic computational information displayed on the interface in the intelligent interaction engine.

[0030] Referring to Figure 5, the ontology behavior logic engine, as the platform's "business behavior definition and logic orchestration center," is based on a unified business operation and business event model. It realizes the abstract modeling of the dynamic behavior of business object ontology and the data-driven definition of business semantics, replacing the existing "hard-coded" implementation of business logic. It defines, visualizes, orchestrates, and manages the business rules and logic processes of business operations and business event behaviors in a "data-driven and model-driven" manner, solving the problem of business logic being closed in a "code black box." This makes it a business model asset that can be "readable, understandable, usable, and optimized" by business personnel and AI machines, achieving transparent management and intelligent execution of business logic.

[0031] Based on a unified behavior model, the semantics, conditions, inputs, process logic, and outputs of business operations and events are defined in a "visualized, structured, and data-driven" manner. It parses and executes the business operation and event models, calls the computational functions or business service capabilities of the ontology function engine, and supports integrated calls to Large Language Model (LLM) services to achieve automated and intelligent execution of business processes. It monitors changes in the state of business operations, system events, or business entity data, and automatically drives the logical flow execution of business operations and events when the operation or triggering conditions of the behavior logic model are met.

[0032] The business operations and business event models defined by the ontology behavior logic are stored in a unified business ontology model. During execution, the business semantics, conditions, inputs, process logic flow, and output rules of the defined ontology behaviors are retrieved from the business ontology model. During the execution of the behavior logic flow, the data services of the ontology data engine are called to obtain the instance data state of the business entity, and the specific calculation functions of the ontology function engine or the execution capability units of the business services are called.

[0033] The intelligent interaction engine intelligently identifies, dynamically acquires, and displays specific business operations of relevant business entities that meet user task requirements. These operations are then triggered by user actions, resulting in the execution of the defined behavioral logic model. During the execution of this behavioral logic flow, business services from external systems accessed by the system integration engine can be invoked, enabling integration with these external systems' business processes. The business rules and logical flow models defined by the ontology behavioral logic engine, along with the execution process and results of their running instances, constitute a "deterministically controlled" chain of business facts. This chain is one of the primary sources of business knowledge for the knowledge generation engine to perform reasoning analysis, discovery verification, and generation optimization of implicit business rules and new knowledge.

[0034] Referring to Figure 6, the system integration engine acts as the "internal and external connector" of the platform. Based on a unified business ontology model, the platform and external systems achieve "two-way, standardized" business service integration and data integration, enabling real-time linkage and collaboration of enterprise-wide business application system processes, and full-domain data association query and reasoning analysis.

[0035] The platform registers and connects to external applications, orchestrates service integration processes based on a unified business service model, and achieves business process integration and real-time linkage between the platform and external systems. It accesses external system data sources, enabling the "association, mapping, and fusion" of external and internal data based on a unified business ontology model to construct a comprehensive data view. Furthermore, it supports automatically generating the platform's business entities and their data models from external data models.

[0036] System integration engine for external data models With business ontology model The entity-relationship mapping uses a joint optimization objective to determine the mapping function. , so that: in: For the set of nodes of the external model, For the set of edges of the external model, The set of nodes in the ontology. The set of edges in the ontology. For the optimal mapping function, This means taking the option that maximizes the objective function. , For nodes in the node set of the external model, Weights for semantic similarity terms. For semantic similarity function, external node semantic representation, This is the semantic representation of the ontology node. and For the ontology node, These are two nodes connected in the external model. For the structural consistency term weight, For the set of ontology relations, For indicator functions, For the penalty intensity weight, For mapping functions Inconsistency measure.

[0037] The "external service contracts, data mapping rules, integration processes, etc." of the system integration model are defined in the unified business ontology model. During execution, the defined system integration model is read from the business ontology model. The "data integration" capability provided by the system integration engine transforms data sources from external systems into internal semantic models, providing the ontology data engine with "new and more comprehensive" business instance data injection, and enhancing the ontology data engine's ability to perform internal and external full-domain data association queries and reasoning analysis based on the unified business ontology model. The "service integration" capability provided by the system integration engine orchestrates and combines internal business services registered in the ontology function engine with external business services accessed by the system integration engine, realizing "two-way" business capability interoperability and process integration between the platform and external systems. Based on the "internal and external connector" capability provided by the system integration engine, the ontology behavior logic engine, intelligent interaction engine, and knowledge generation engine can simultaneously acquire and operate information and capabilities of the platform and external systems, enriching and expanding the application scenarios of "behavioral linkage, system interaction, and knowledge creation" based on the unified business ontology model.

[0038] Referring to Figure 7, the intelligent interaction engine, as the platform's "user interface generator and intelligent interaction agent," drives the automatic generation of page models for user interaction interfaces based on the business ontology model, and understands the user's natural language task requirements, dynamically assembling and presenting the interactive interfaces and functions required for the task, thereby realizing an intelligent interaction mode of "task requirement-driven and function proactively adapting to people."

[0039] Based on the business ontology model, UI mapping framework, and page component library, metadata models such as "business entities, attributes, relationships, and business operations" are automatically mapped and generated into "page components, page fields, page layouts, and operation buttons" for the user interface. The system understands the user's natural language task requirements, automatically identifies relevant business entities and operations, and the required page components. In this embodiment, a component set is selected from the page component library to dynamically assemble the task interaction interface, satisfying the coverage constraints of the task requirements and optimizing the objective function. in: This indicates finding the minimum value. Select a vector for the components, indicating which components to choose from the component library to enter the current task page. , Each dimension can only take the value 0 or 1, and there are a total of dimension, This represents the total number of components in the page component library. For the semantic requirements set of the current user task, for One of the specific requirements in the document. For demand Importance weights For component indexing, For the first Binary selection variables for each component, For the first Candidate components Representation Component Demand coverage The weighting coefficient for the interaction cost item. For components Interaction cost The weighting coefficient for redundant terms. This indicates the relationship between all different components. Summation of combinations, For the first Binary selection variables for each component, For components and Conflict / redundancy coefficient.

[0040] Then, based on the page model configuration framework and interaction engine requirements, a personalized task interaction interface is dynamically assembled and presented. A unified page component registration model and integration framework are provided, supporting the registration of user-developed customized pages as platform-runnable page components. Based on a unified page component integration standard, the development results of various application system projects can be fed back into the platform's page component library for reuse.

[0041] The UI mapping rules upon which the intelligent interaction engine's page model generation depends are stored in a unified business ontology model. When automatically generating or dynamically assembling the user interface, it retrieves the UI mapping rules and business semantics from the business ontology model. For a given user task, the intelligent interaction engine acts as the overall coordinator for the user: it retrieves and updates data content from the ontology data engine, retrieves and executes business operations from the ontology behavior logic engine, and finally dynamically assembles the data into a page layout and presents the user interface.

[0042] User commands issued through the intelligent interaction engine are the source of the entire platform (ontology data query, ontology behavior logic execution, and internal and external business service integration calls). User interaction behavior data is recorded through the intelligent interaction engine and fed back to the business ontology model and knowledge generation engine for the platform to learn and optimize, and to automatically evolve towards a more personalized and dynamic user interaction experience.

[0043] Referring to Figure 8, the knowledge generation engine, acting as the platform's "intelligent brain," utilizes the knowledge generalization reasoning and generation capabilities of the Large Language Model (LLM) based on specific prompt word templates and user natural language task scenario requirements. It combines the deterministic business knowledge of the business ontology model and instance data with the built-in professional algorithm library for industrial fields as contextual input for LLM reasoning analysis. This automatically discovers implicit business knowledge, generates corresponding knowledge service components, and writes them back to the unified business ontology model, automatically expanding them into new capability units of the business ontology, continuously driving the evolution of business knowledge and cognitive evolution.

[0044] From the business ontology model and its business instance data, structured "business entity relationships, attribute constraints, business behaviors, business rules, and logical processes" are automatically extracted as known deterministic knowledge. Specific prompt word templates are constructed, and based on user task scenario requirements analysis, the large language model is guided to automatically match and combine deterministic business knowledge with a built-in industry-specific algorithm library to perform causal reasoning, pattern discovery, and hypothesis generation. This automatically discovers new business patterns and implicit knowledge, and automatically generates knowledge solutions.

[0045] The knowledge solutions discovered by the large language model inference are tested and verified online based on the operational instance data of the business ontology or the simulation data generated by the large language model. The verified knowledge solutions are selected and automatically published and registered as knowledge service components (calculation functions, business rules, business services and page components, etc.) that can run on the platform, and automatically updated to the business ontology model, forming a "cognition-application" closed loop of the business ontology.

[0046] Deterministic business knowledge is extracted from the unified business ontology model. Newly generated knowledge service components are used as extensions of the business ontology model and written back to the unified business ontology model, completing the knowledge loop. Business entity instance data is obtained from the ontology data engine, and the "business rules and logical processes" model and execution logs of business behaviors are obtained from the ontology behavior logic engine for fusion reasoning analysis. The knowledge generation engine produces "new or optimized" computation functions and business services, continuously expanding and optimizing the data state observation and change capabilities of the business ontology. The "new or optimized" business rules and logical processes produced by the knowledge generation engine are injected into the ontology behavior logic engine, changing the future execution logic of business behaviors.

[0047] The knowledge generation engine continuously generates new business knowledge by learning the business ontology model and runtime instance data ("data" + "logic" + "interaction") generated by the continuous learning platform. This continuously expands and optimizes the platform's future behavioral capabilities and behavioral logic, forming a complete cognitive evolution loop of "running data - knowledge discovery - knowledge generation - knowledge application - knowledge optimization".

[0048] This embodiment also applies the intelligent business application platform based on ontology model and engine drive to three aspects: product design and manufacturing collaboration, production process and quality optimization, and predictive maintenance and fault diagnosis of equipment.

[0049] Scenario 1: Product Design and Manufacturing Collaboration The following section uses the typical industrial scenario of "product design and manufacturing collaboration" as an example to illustrate the specific implementation of this invention. In this scenario, when product design changes (such as adjustments to structure, materials, or process parameters), the change information needs to be accurately, automatically, and synchronously integrated into downstream processes such as process design, production planning, material supply, and manufacturing execution. Traditional methods rely on manual communication and cross-system verification, resulting in low collaboration efficiency and inconsistencies. By constructing a unified business ontology model and relying on a multi-drive engine collaborative working mechanism, the automatic identification of change impacts, intelligent task assignment, process-assisted decision-making, and closed-loop error prevention on-site are achieved, ultimately forming a continuously optimized knowledge accumulation system.

[0050] The specific implementation steps are as follows: Step 1: Construct a unified business ontology model. Organize experts in design, process, manufacturing, and supply chain to jointly construct a domain-specific business ontology model covering the entire product lifecycle, using the system's provided visual modeling tools. This model clearly defines core entities (such as parts, processes, work orders, orders, and change orders) and their attributes, establishes semantic relationships between entities (such as "parts - used for - work orders" and "process - consumed - materials"), and sets business constraint rules (such as technical parameter ranges and compliance requirements). This model serves as a unified, machine-understandable semantic foundation for the entire process, ensuring the consistency and accuracy of information interaction across systems and stages.

[0051] Step 2: Standardized capture and semantic processing of design change events. After the upstream Product Design Management (PLM) application completes the change approval, it pushes the structured change event to the platform through the standardized business service interface provided by the system integration engine. The event content includes the change object identifier, version evolution information, and specific change details.

[0052] After receiving an event, the ontology data engine automatically creates or updates relevant entity objects (such as new version components) based on the unified business ontology model, refreshes their attributes and relationships, and constructs a complete version evolution and relationship chain in the dynamic knowledge graph, providing a real-time and accurate semantic data foundation for subsequent impact analysis.

[0053] Step 3: Automatic Analysis of Change Impact Based on Dynamic Knowledge Graph. The data engine, based on a constructed dynamic knowledge graph linking design, process, resources, and planning elements, performs graph traversal and inference calculations to automatically and in real-time analyze the scope of the changes. The system accurately identifies affected work-in-process and pending production tasks, process documents requiring synchronized updates, procurement or outsourcing orders requiring adjustment, and manufacturing resources such as tooling and molds that need re-verification or preparation.

[0054] Based on the impact analysis results and combined with the predefined "role-task" mapping rules, the intelligent interaction engine dynamically generates and pushes personalized work lists and dashboards to relevant process engineers, production planners, purchasing specialists, etc., highlighting pending tasks and recommending action plans.

[0055] Step 4: Data and Model-Driven Process Collaboration and Decision Support. Affected personnel (such as process engineers) initiate a collaborative workflow upon receiving the task. The ontology behavior logic engine drives predefined collaborative workflow templates, automatically forming cross-functional virtual collaboration teams and assigning tasks.

[0056] The knowledge generation engine is activated during this process, providing multi-dimensional decision support: extracting the associated knowledge from the business ontology model and the runtime instance database; driving analytical models (such as large language models) to parse the changes and generating specific process adjustment suggestions based on historical experience data; and verifying the feasibility of the preliminary solution through integrated simulation services. After cross-departmental online review and optimization, the finalized process data is automatically synchronized to the downstream Manufacturing Execution System (MES) through the system integration engine.

[0057] Step 5: Visualized error prevention and execution feedback on the manufacturing floor. When a production task involving changes arrives at the manufacturing unit, the intelligent interaction engine dynamically pushes augmented reality guidance interfaces or mobile operation instructions based on the operator's role and workstation configuration, clearly indicating the correct material version, process parameters, and operating steps to be used.

[0058] Operators scan material or vehicle identification tags, and the system verifies the consistency between the physical item and the task requirements in real time. If a version or specification mismatch is found, the system immediately issues a warning and can restrict subsequent operations according to rules to prevent the error from continuing.

[0059] Real-time process data from on-site intelligent terminals (such as sensors and CNC equipment) is collected synchronously, and the behavioral logic engine performs online anomaly detection based on the quality rule base: qualified data is automatically associated and archived; abnormal data triggers a predefined quality handling process and notifies relevant personnel to intervene and analyze.

[0060] Step 6: Closed-loop learning and automated knowledge accumulation. The system fully collects and correlates all data from the entire process of this change, from "initiation, decision-making, execution to verification".

[0061] The knowledge generation engine explores the implicit patterns between process parameters, manufacturing processes and final product quality through periodic data mining and machine learning analysis (such as using large language models for text and data association analysis), forming empirical rules such as "when using a new material, recommending a certain range of key process parameters can achieve better performance".

[0062] These rules are automatically transformed into structured knowledge entries and registered as new business constraints or optimization suggestions in the business rule base of the unified business ontology model, thereby empowering subsequent change decisions and process planning and achieving the co-evolution of system knowledge and business capabilities.

[0063] By applying this implementation method, the design change and manufacturing collaboration process is significantly improved: the time for change information to permeate the entire process is reduced from weeks to days. Real-time verification and on-site error prevention mechanisms based on semantic consistency effectively eliminate quality defects caused by information mismatch. The approach shifts from relying on fragmented individual experience to collaborative decision-making based on unified data assets and model support. The experience gained from each change handling is transformed into structured knowledge, continuously feeding back into the business ontology and rule base, driving continuous optimization of the management system.

[0064] Scenario 2: Production Process and Quality Optimization The following section uses the typical industrial scenario of "production process and quality optimization" as an example to explain the specific implementation of this invention in detail. In this scenario, traditional methods rely on manual experience for setting process parameters and conducting post-event quality analysis, which suffers from problems such as slow response, long optimization cycles, and difficulty in knowledge accumulation. This invention constructs a unified business ontology model and coordinates multiple driving engines to achieve semantic interconnection of all production elements, intelligent prediction of quality risks, proactive collaborative execution of optimization measures, and closed-loop learning and self-evolution of process knowledge.

[0065] Specific implementation steps: Step 1: System initialization and unified business modeling. Construct a unified business ontology model: Experts in process, equipment, quality, and other fields utilize the platform's visual semantic modeling tools to jointly define the core business entities and their relationships within the manufacturing domain. The business ontology model clearly defines entities such as "production equipment," "process formula," "production task," and "quality characteristics," along with their attributes (e.g., equipment status parameters, process parameter values, quality indicators), and defines semantic relationships between entities (e.g., "production task - adopting - process formula," "production equipment - output - product (associated with quality characteristics)"). This model serves as the semantic benchmark for the entire system to consistently understand business data.

[0066] Define business logic and interaction rules: Based on the above ontology model, key business process logic is defined through the behavior logic engine, such as the "process parameter optimization and adjustment process," clarifying its triggering conditions, judgment logic, execution steps, and inter-system call relationships. Simultaneously, user interface mapping rules are configured through the intelligent interaction engine to dynamically map specific business entities and their data states to corresponding visual page components, laying the foundation for subsequent personalized task interface assembly.

[0067] Step 2: Multi-source data fusion and dynamic knowledge graph construction. An integration engine connects data sources such as manufacturing execution applications, equipment monitoring applications, and quality management applications to achieve automatic collection of heterogeneous data, including production tasks, real-time equipment operation data, historical process parameters, and quality inspection results.

[0068] Based on the unified business ontology model defined in step one, the ontology data engine performs semantic annotation and association on the input data, and builds and updates a dynamic knowledge graph covering the entire domain of "equipment-process-production-quality" in real time in the background. This graph dynamically presents the real-time association status between various production elements in a graphical way, so that "an anomaly of a certain equipment parameter" can be immediately associated with "the production task being executed", "the process parameters used", and "the quality results under similar historical conditions".

[0069] Step 3: Data- and Model-Based Intelligent Risk Prediction. Implicit Knowledge Mining: The knowledge generation engine periodically extracts data from dynamic knowledge graphs and historical instance databases, combining this with semantic constraints provided by a unified business ontology model to construct analytical tasks. For example, it drives analytical models (such as large language models) to explore potential correlation patterns between equipment operating parameters, process parameters, and key quality indicators.

[0070] Automated generation and deployment of predictive rules: Analysis models may discover potential patterns such as "when parameter A is in interval X and parameter B shows trend Y, the probability of quality indicator Z exceeding the standard in the subsequent N production units increases significantly." The knowledge generation engine automatically transforms such findings into structured, machine-executable predictive or early warning rules (IF-THEN form) and registers them as new knowledge assets in the business rule base of the unified business ontology model. The behavioral logic engine synchronously loads these new rules, making them effective for real-time monitoring.

[0071] Step 4: Proactive Intervention and Cross-System Intelligent Collaboration. Risk Warning and Contextualized Task Interface Generation: When real-time data triggers warning rules, the behavioral logic engine generates a warning event. The intelligent interaction engine dynamically assembles a handling workbench for specific roles (such as process engineers and equipment maintenance personnel) based on the event type, affected entities, and predefined UI rules. This interface integrates real-time data visualization, historical case comparison, related knowledge prompts, and predefined business operation entry points (such as "Start Parameter Optimization" and "Initiate Maintenance Request"), aggregating discrete information and functions into a unified decision-making context.

[0072] Automated process execution: Users trigger predefined business operations (such as "execute process optimization") on the interface. The ontology behavior logic engine drives the corresponding business logic model to automatically execute a series of actions: calling the optimization algorithm service registered in the ontology function engine to calculate recommended parameters; calling the business service interface of external systems through the system integration engine to automatically initiate the necessary approval process; after approval, automatically sending the approved optimization parameters to the production site control system to complete the closed-loop adjustment of process parameters.

[0073] Step 5: Effect Verification and Closed-Loop Knowledge Learning. Data Association and Accumulation Throughout the Process: All data from the entire process—from risk prediction and decision intervention to execution adjustments—including triggering conditions, intervention measures, adjusted process parameters, and final product quality results, are fully captured and associated with the ontology data engine and recorded in a dynamic knowledge graph, forming a traceable "decision-effect" fact chain.

[0074] Knowledge Validation and Model Iteration: In subsequent analysis cycles, the knowledge generation engine will validate and iteratively optimize the generated business rules or data models based on updated data, including the results of this intervention. For example, it may confirm the validity of the original rules or discover better combinations of intervention parameters, thereby generating more accurate business rules or optimizing the parameters of existing algorithm calculation function models. In this way, the system achieves a continuous self-evolutionary cycle from "data" to "knowledge," and then uses "knowledge" to optimize "decision and execution."

[0075] Through the application of this implementation method, the production process and quality management model achieve the following transformation: from "post-event inspection and trial-and-error" to "pre-event prediction and proactive optimization." Through semantically consistent information exchange and process automation, rapid and accurate collaboration across personnel roles and business systems is achieved. Decision support shifts from relying on "human single-point historical experience" to automated and intelligent reasoning analysis based on "cross-domain business data association." Implicit expert experience rules and business data patterns are automatically discovered and transformed into structured, reusable, and iterative business knowledge assets, driving continuous self-optimization of the production system.

[0076] Scenario 3: Predictive Equipment Maintenance and Fault Diagnosis The following section uses the typical industrial scenario of "predictive equipment maintenance and fault diagnosis" as an example to illustrate the specific implementation of this invention. Traditional maintenance models often rely on periodic inspections or reactive repairs, resulting in high unplanned downtime rates, high maintenance costs, and difficulties in knowledge transfer. This invention constructs a unified equipment operation and maintenance semantic model, coordinating multiple driving engines to achieve real-time perception of equipment status, early prediction of potential faults, intelligent assistance in diagnostic decisions, and precise closed-loop maintenance execution, while driving continuous self-evolution of equipment operation and maintenance knowledge.

[0077] Specific implementation steps: Step 1: Construct a unified business ontology model. Equipment management, maintenance experts, and process personnel will use the visual semantic modeling tools provided by the platform to jointly define the core business concept system in the field of equipment operation and maintenance: Core entity definition: including business entities such as "production equipment", "core components", "sensors", "failure modes", "maintenance work orders", "spare parts materials", and "production tasks", and clarify their attributes (such as equipment technical parameters, sensor range, and fault characteristic values).

[0078] Semantic relationship definition: Establish the association between business entities, such as "equipment-contains-component", "component-assembly-sensor", "equipment-execution-production task", "failure mode-corresponding-maintenance procedure", "maintenance procedure-consumption-spare parts".

[0079] Constraints and rules definition: Set equipment health status thresholds (such as vibration intensity limits, temperature alarm thresholds), maintenance procedures and standards, etc.

[0080] This model provides a consistent semantic understanding of devices, data, processes, and knowledge across the entire system.

[0081] Step 2: Multi-source heterogeneous data fusion and dynamic knowledge graph construction. Connect and access various data sources through the system integration engine: Real-time runtime data: Access time-series data streams from device sensors (such as vibration, current, temperature, and pressure) from device monitoring applications or IoT platforms.

[0082] Production context data: Obtain production tasks, process parameters, and product information associated with equipment from manufacturing execution applications.

[0083] Static master data and historical data: Obtain equipment material lists, technical documents, and maintenance records from asset management applications.

[0084] Environmental and operating condition data: Access relevant environmental parameters (temperature, humidity) and energy data.

[0085] Based on the business ontology model defined in step 1, the ontology data engine performs real-time semantic annotation, alignment, and association on the aforementioned multi-source data, automatically constructing and continuously updating a dynamic knowledge graph covering the entire domain of "equipment-component-operating condition-environment-maintenance history" in the background. This graph enables real-time semantic interconnection between equipment status, production activities, and historical experience.

[0086] Step 3: Data and Model-Based Fault Prediction and Diagnosis Knowledge Generation. The knowledge generation engine initiates periodic analysis tasks to extract information from dynamic knowledge graphs and long-term historical data.

[0087] Latent pattern mining: The engine combines the semantic framework provided by the business ontology to drive the analysis model (such as the large language model) to explore the complex correlation and early symptom patterns between equipment operating parameters, process load, environmental factors and specific failure modes.

[0088] Automated construction of executable knowledge: Implicit business patterns discovered through analysis (e.g., "When a certain type of equipment is performing a high-load specific process and auxiliary system parameter A is abnormal, the probability of component B failing within N hours increases significantly") are automatically transformed into structured executable knowledge: predictive early warning rules: clearly define triggering conditions, risk levels, and recommended inspection times.

[0089] Assisted diagnostic recommendations: List potential causes, recommended investigation steps, and required resources.

[0090] The newly generated knowledge is automatically registered as an asset to the business rule base of the unified business ontology model and synchronized to the ontology behavior logic engine and equipment maintenance management application.

[0091] Step 4: Intelligent Early Warning Triggering and Maintenance Task Coordination Assignment. When the real-time data stream meets a certain early warning rule condition, the ontology behavior logic engine is automatically triggered to execute the collaborative process: Multi-channel early warning release: Visual alarms are displayed on the centralized monitoring dashboard, and notifications are pushed to relevant responsible personnel (equipment administrators, maintenance engineers, etc.) through collaborative tools.

[0092] Intelligent maintenance work order creation: Preventive maintenance or inspection work orders are automatically generated in equipment maintenance management applications. The work orders are associated with diagnostic suggestions, required tools and skill requirements, and can be intelligently recommended or assigned based on personnel location, skills and workload.

[0093] Resource preparation: Based on diagnostic suggestions, automatically check the inventory status of possible spare parts and generate an alert when inventory is insufficient.

[0094] Step 5: Intelligent assistance and execution feedback at the maintenance site. Maintenance personnel receive tasks via mobile terminals or augmented reality devices.

[0095] Contextualized work instructions: The intelligent interaction engine dynamically generates and pushes enhanced work instructions based on the work order content, equipment model, and maintenance personnel role. These instructions may include: overlaying the 3D model of the equipment with the physical location, highlighting key points of the standard operating procedures, linking historical cases and knowledge, and reminding users of safety regulations.

[0096] Digital operation execution: Maintenance personnel perform inspection, testing, or replacement operations according to instructions, and record process data (such as measured values ​​and replacement part information) and confirm the completion of steps through a terminal. Post-maintenance equipment trial operation data is automatically collected.

[0097] Step 6: Closed-loop feedback and continuous knowledge optimization. The entire process of early warning, diagnosis, repair, and verification is fully recorded, forming a closed-loop business fact chain of "prediction-intervention-result," and is stored in a dynamic knowledge graph.

[0098] In subsequent analysis cycles, the knowledge generation engine utilizes newly accumulated closed-loop data to validate and iterate the effectiveness of deployed early warning rules and diagnostic logic. For example, it may revise early warning thresholds to improve prediction accuracy or timeliness, enrich the diagnostic tree by adding new fault characteristic identifiers, and output optimization suggestions for equipment operation or process parameters to reduce the risk of failure at its root. Through this closed loop, the equipment operation and maintenance knowledge base and decision-making model can continuously improve themselves.

[0099] This implementation method enables a systematic improvement in equipment operation and maintenance management: it shifts from reactive maintenance to predictive maintenance, significantly reducing unplanned downtime and improving overall equipment efficiency. Condition-based precision maintenance avoids over-maintenance, and predictive spare parts demand alerts help optimize inventory costs. Intelligent guidance for on-site repairs shortens average diagnostic and repair times and promotes standardized maintenance procedures. Expert experience and data patterns are transformed into reusable and iterative structured knowledge, enabling effective knowledge accumulation and transfer. Maintenance tasks can be intelligently coordinated with production plans, minimizing the impact on production activities.

[0100] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects have an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0101] The above description is merely a specific embodiment of this application. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. The protection scope of this application should be determined by the protection scope of the claims.

Claims

1. An intelligent business application platform based on ontology model and engine-driven architecture, characterized in that, include: The business ontology model includes an ontology data engine, an ontology function engine, an ontology behavior logic engine, a system integration engine, an intelligent interaction engine, and a knowledge generation engine that interact with and are connected to the business ontology model. The business ontology model serves as the sole trusted source and information coordination hub, and interacts and collaborates with the aforementioned engines through a pre-defined internal interface based on a unified semantic model. The ontology data engine provides integrated and semantically consistent data services to all engines, generating a knowledge graph of the core context. The ontology function engine provides atomic computing and service capabilities, which are called by the ontology behavior logic engine to execute business processes, and orchestrated by the system integration engine to provide data and business services to the outside world. The ontology behavior logic engine models business processes. Its defined process model can be triggered by the intelligent interaction engine to generate user operations, and can also be analyzed and optimized by the knowledge generation engine. The intelligent interaction engine directly provides users with a dynamic user interface driven by task requirements. It uses the ontology data engine to acquire data, calls the ontology behavior logic engine to execute business operations, and enriches the interaction context based on external information accessed by the system integration engine. The knowledge generation engine continuously learns from the full-domain operation data, behavior logic execution records, and interaction feedback, and feeds the new knowledge it generates back to the business ontology model, ontology function engine, ontology behavior logic engine, and intelligent interaction engine to continuously optimize the business ontology model and drive the continuous evolution of the system. The system integration engine ensures that the business capabilities and data of external systems can be integrated for data exchange and collaboration.

2. The intelligent business application platform based on ontology model and engine-driven as described in claim 1, characterized in that, The business ontology model serves as the sole trusted source and information coordination hub. It interacts and collaborates with the aforementioned engines through a pre-defined internal interface based on a unified semantic model. This includes: the business ontology model receiving and updating business entity instance data and relationships written by the ontology data engine; the ontology function engine reading and converting the state of the business entity data written by the business ontology model; the business ontology model receiving execution fact chains from the ontology behavior logic engine, which reads the logical definitions of business operations or events from the business ontology model and writes them into execution instances; the business ontology model receiving updates to the mapping between external data sources and instance data accessed by the system integration engine, which reads the integration contract from the business ontology model and converts it into the state of the business entity data; the intelligent interaction engine obtaining the business semantic context of user task understanding from the business ontology model; the business ontology model receiving newly discovered knowledge and rules from the knowledge generation engine and storing them as continuously evolving business knowledge memory for the platform; and the knowledge generation engine using the business ontology model as a deterministic knowledge source to automatically reason, generate, and write the newly discovered knowledge.

3. The intelligent business application platform based on ontology model and engine-driven as described in claim 1 or 2, characterized in that, The ontology data engine provides integrated and semantically consistent data services to all engines, generating a knowledge graph of the core context, including: data modeling and graph generation based on the semantic definition of the unified business ontology model; providing raw data for the computation functions of the ontology function engine, providing context data after business semantic fusion for the logic execution of the ontology behavior logic engine, providing display data after business semantic fusion for the interface presentation of the intelligent interaction engine, and providing structured instance data after business semantic fusion for the analysis and reasoning of the knowledge generation engine; receiving external data processed by the data integration part of the system integration engine and writing it into the business ontology model and the unified data view integrated into the ontology.

4. The intelligent business application platform based on ontology model and engine-driven as described in claim 3, characterized in that, Before writing multi-source data into the business ontology model, the ontology data engine constructs a semantic consistency fusion objective and minimizes the fusion loss: in: The number of data sources participating in the integration, For the first The credibility weight of each data source The final attribute values ​​after fusion, i.e., the reliable results prepared for writing into the business ontology model, For the first Observations of the same business entity attribute from multiple data sources. As a measure of difference, it measures the cost of the fused value deviating from the observations of the data source. These are the weighting coefficients for semantic consistency constraints. Represents the set of constraints on the ontology. Summing each constraint instance in the table, As the main entity, and For two object entities related to the subject, For the ontology constraint set, This is a function that maps entities to graph embedding vectors. It is the square of the L2 norm.

5. The intelligent business application platform based on ontology model and engine-driven as described in claim 3 or 4, characterized in that, The ontology function engine provides atomic computation and service capabilities, which are invoked by the ontology behavior logic engine to execute business processes, and simultaneously orchestrated by the system integration engine to provide data and business services to external entities. This includes: the ontology function engine defining computation functions and business service models, registering and managing them in a unified business ontology model to ensure semantic consistency; these computation functions and business services being the basic execution capability units for orchestration and invocation by the ontology behavior logic engine; when an ontology function is invoked and executed, it obtains the input or output parameter rules and business semantics of the ontology function from the business ontology model; the business services registered by the ontology function engine are discovered, orchestrated, and integrated with the business processes of external systems by the service integration component of the system integration engine; and it supports the backend capabilities of the intelligent interaction engine, realizing the specific backend functions of the behavioral logic triggered by user operations or the dynamic computational information displayed on the interface within the intelligent interaction engine.

6. The intelligent business application platform based on ontology model and engine-driven as described in claim 5, characterized in that, The ontology behavior logic engine models the business process, including: visually, structurally, and data-drivenly defining the semantics, conditions, inputs, process logic flow, and outputs of business operations and business events based on a unified behavior model; parsing and executing the business operation and business event model, calling the computation functions or business service capabilities of the ontology function engine, and supporting the integration and invocation of large language model services to achieve automated and intelligent execution of business processes; monitoring changes in the status of business operations, system events, or business entity data, and automatically driving the execution of the logical flow of business operations and business events when the operation or triggering conditions of the behavior logic model are met.

7. The intelligent business application platform based on ontology model and engine-driven as described in claim 1, characterized in that, The system integration engine ensures that the business capabilities and data of external systems can be integrated for data exchange and collaboration, including: registering and connecting external applications, orchestrating service integration processes based on a unified business service model, realizing business process integration and real-time linkage between the platform and external systems; the platform accesses external system data sources, enabling external data and internal data to be associated, mapped and merged based on a unified business ontology model to build a full-domain data view, and supporting the automatic reverse generation of the platform's business entities and their data models based on external C data models.

8. The intelligent business application platform based on ontology model and engine-driven as described in claim 7, characterized in that, System integration engine for external data models With business ontology model The entity-relationship mapping uses a joint optimization objective to determine the mapping function. , so that: in: For the set of nodes of the external model, For the set of edges of the external model, The set of nodes in the ontology. The set of edges in the ontology. For the optimal mapping function, This means taking the option that maximizes the objective function. ， For nodes in the node set of the external model, Weights for semantic similarity terms. For semantic similarity function, external node semantic representation, This is the semantic representation of the ontology node. and For the ontology node, These are two nodes connected in the external model. For the structural consistency term weight, For the set of ontology relations, For indicator functions, For the penalty intensity weight, For mapping functions Inconsistency measure.

9. The intelligent business application platform based on ontology model and engine-driven as described in claim 1, characterized in that, The intelligent interaction engine directly provides users with a dynamic user interface driven by task requirements. This includes: automatically mapping business entities, attributes, relationships, and business operation metadata models to generate page components, page fields, page layouts, and operation buttons for the user interface based on the business ontology model, UI mapping framework, and page component library; parsing user natural language task requirements, automatically identifying relevant business entities and operations, as well as the required page components, and selecting a component set from the page component library to dynamically assemble the task interaction interface, satisfying the coverage constraints of task requirements and optimizing the objective function. in: This indicates finding the minimum value. Select a vector for the components, indicating which components to choose from the component library to enter the current task page. ， Each dimension can only take the value 0 or 1, and there are a total of dimension, This represents the total number of components in the page component library. For the semantic requirements set of the current user task, for One of the specific requirements in the document. For demand Importance weights For component indexing, For the first Binary selection variables for each component, For the first Candidate components Representation Component Demand coverage The weighting coefficient for the interaction cost item. For components Interaction cost The weighting coefficient for redundant terms. This indicates the relationship between all different components. Summation of combinations, For the first Binary selection variables for each component, For components and The system calculates the conflict / redundancy coefficients; then, based on the requirements of the page model configuration framework and interaction engine, it dynamically assembles and presents personalized task interaction interfaces; it provides a unified page component registration model and integration framework, supporting the registration of user-developed customized pages as platform-runnable page components.

10. The intelligent business application platform based on ontology model and engine-driven as described in claim 1, characterized in that, The knowledge generation engine continuously learns from full-domain operational data, behavioral logic execution records, and interactive feedback, and feeds the generated new knowledge back to the business ontology model, ontology function engine, ontology behavioral logic engine, and intelligent interaction engine. This continuously optimizes the business ontology model and drives the system's continuous evolution. This includes: automatically extracting structured business entity relationships, attribute constraints, business behaviors, business rules, and logical processes as known deterministic knowledge from the business ontology model and its business instance data; constructing specific prompt word templates, and based on user task scenario requirements analysis, guiding the large language model to automatically match and combine deterministic business knowledge with the built-in industry-specific professional algorithm library to perform causal reasoning, pattern discovery, and hypothesis generation, automatically discovering new business rules and implicit knowledge, and automatically generating knowledge solutions; and conducting online debugging and verification of the knowledge solutions discovered by the large language model based on the operational instance data of the business ontology or the simulation data generated by the large language model. Selecting verified knowledge solutions, automatically publishing and registering them as platform-runnable knowledge service components, and automatically updating them to the business ontology model, forming a closed loop of cognitive application of the business ontology.

Images