AI model combinatorial optimization-based AI business process automatic generation method

Abstract

The invention discloses an AI business process automatic generation method based on AI model combinatorial optimization, and the method comprises the following steps: constructing a multi-level AI capability decoupling and reconstruction module, and carrying out the bottom-up hierarchical modeling and top-down modular decoupling, dynamic mapping of an AI atomic power layer, an AI modular production capacity layer, an AI general capability layer and an AI application business layer is realized; an elastic AI capability combinatorial optimization module is constructed, AI capability combinatorial optimization oriented to three dimensions of data, features and models is carried out based on service quality requirements, and multiplexing, combination and arrangement of AI capabilities are realized through evolutionary computation optimization driven by an agent model; and an AI business process automatic generation module is constructed, an AI service containerization deployment scheme is generated and optimized based on data-driven process mining and a hyper-heuristic algorithm, and AI business process automatic generation is realized. According to the invention, the adaptability and execution efficiency of the AI technology in a complex scene can be improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of business process generation, and in particular to an AI business process automatic generation method based on AI model combination optimization. BACKGROUND

[0002] With the rapid development of artificial intelligence technology, the new infrastructure layout takes the wisdom network national new generation artificial intelligence open innovation platform, the intelligent platform and the digital employee platform as the core, aiming to realize the comprehensive improvement of connection, computing power and ability; however, in the dynamic complex scene, the AI ability of the digital employee still faces the technical bottlenecks of insufficient reusability, low AI component service quality in differentiated environment and poor adaptability in dynamic environment, which directly leads to low execution success rate and limits the wide application of AI technology in actual business.

[0003] Therefore, aiming at the above problems, the combination optimization method of AI ability and the automatic processing technology of business process can effectively improve the adaptability and service quality of AI ability, which has important technical value and application significance. SUMMARY

[0004] In order to overcome the defects and deficiencies existing in the prior art, the application provides an AI business process automatic generation method based on AI model combination optimization, which realizes AI capability decoupling and hierarchical modeling, AI capability on-demand elastic combination optimization and AI business process automation in a dynamic open environment, based on the traction of actual business demand, empowers the digital employee platform, and further promotes the efficient supply of AI ability; a diversified AI ability decomposition and reconstruction is constructed, and the AI ability adaptive combination optimization provides a systematic solution for AI application in a dynamic open environment, so as to improve the adaptability and execution efficiency of AI technology in complex scenes.

[0005] In order to achieve the above purpose, the application adopts the following technical scheme:

[0006] The application provides an AI business process automatic generation method based on AI model combination optimization, comprising the following steps:

[0007] A multi-level AI capability decoupling and reconstruction module is constructed, which realizes the dynamic mapping of AI atomic capability layer, AI modular production capability layer, AI general capability layer and AI application business layer through hierarchical modeling from bottom to top and modular decoupling from top to bottom;

[0008] An elastic AI capability combination optimization module is constructed, which realizes AI capability combination optimization in three dimensions of data, features and models based on service quality demand, and realizes AI capability reuse, combination and arrangement through proxy model driven evolutionary calculation optimization;

[0009] An AI business process automation generation module is constructed, based on data-driven process mining and super-heuristic algorithms, to generate and optimize AI service containerization deployment schemes, and realize AI business process automation generation.

[0010] As a preferred technical solution, the AI atomic capability layer is used to realize atomic-level optimization of AI capabilities.

[0011] The AI modular production capability layer is used to generate modular AI capability components and reusable AI components.

[0012] The AI general capability layer is used to assemble AI components based on business requirements to form functional units.

[0013] The AI application business layer is used to map functional units to business processes through super-heuristic algorithms.

[0014] As a preferred technical solution, through proxy model driven evolutionary computation optimization, the reuse, combination and arrangement of AI capabilities are realized, specifically including:

[0015] A multi-dimensional QoS evaluation proxy model is constructed by superimposing RBF kernel functions and linear kernel functions to evaluate the performance of AI capability combination in QoS.

[0016] An evolutionary computation method is used to generate a candidate model population, and a hybrid mechanism of proxy evaluation and real evaluation is used to update the model library to complete the update and combination of candidate models.

[0017] As a preferred technical solution, based on data-driven process mining and super-heuristic algorithms, AI service containerization deployment schemes are generated and optimized, specifically including:

[0018] High-order features of historical business processes are extracted through process mining technology.

[0019] Genetic programming method is used to learn super-heuristic rules to generate containerization deployment schemes.

[0020] The application also provides an AI business process automation generation system based on AI model combination optimization, which is used to realize the AI business process automation generation method based on AI model combination optimization, and includes a multi-level AI capability decoupling and reconstruction module, an elastic AI capability combination optimization module, and an AI business process automation generation module.

[0021] The multi-level AI capability decoupling and reconstruction module is used to realize dynamic mapping of the AI atomic capability layer, the AI modular production capability layer, the AI general capability layer and the AI application business layer through bottom-up hierarchical modeling and top-down modular decoupling.

[0022] The elastic AI capability combination optimization module is used for AI capability combination optimization in three dimensions of data, features and models based on service quality requirements, and AI capability reuse, combination and arrangement are realized through proxy model driven evolutionary computation optimization.

[0023] The AI business process automation generation module is used for generating and optimizing AI service container deployment schemes based on data-driven process mining and super-heuristic algorithms, and realizing AI business process automation generation.

[0024] As a preferred technical solution, the elastic AI capability combination optimization module comprises an evolutionary computation optimization unit, a QoS evaluation unit and a model library management unit.

[0025] The evolutionary computation optimization unit is used for realizing AI capability combination evaluation and optimization through proxy model driving.

[0026] The QoS evaluation unit is used for QoS evaluation of AI capability combination performance based on a proxy model.

[0027] The model library management unit is used for generating a candidate model population by using an evolutionary computation method, updating a model library by combining a hybrid mechanism of proxy evaluation and real evaluation, and completing candidate model updating and combination.

[0028] As a preferred technical solution, the AI business process automation generation module comprises a process mining unit, a process optimization unit and a containerization deployment unit.

[0029] The process mining unit is used for business process transparency and twinning based on log data, operation data and basic data.

[0030] The process optimization unit is used for mining business logic and generating process optimization rules based on a super-heuristic algorithm.

[0031] The containerization deployment unit is used for generating a containerization deployment scheme.

[0032] As a preferred technical solution, the process optimization unit is used for mining business logic and generating process optimization rules based on a super-heuristic algorithm, and specifically comprises:

[0033] A reference process model pool is constructed based on process twinning technology, and an adaptability function is designed based on QoS measurement.

[0034] AI services are encapsulated based on containerization technology, and deployment strategies are adjusted according to real-time resource states.

[0035] The application further provides a computer readable storage medium storing a program, which is executed by a processor to realize the AI business process automation generation method based on AI model combination optimization.

[0036] The application also provides a computer device comprising a processor and a memory for storing a program executable by the processor, wherein the processor implements the AI service process automatic generation method based on AI model combination optimization when executing the program stored in the memory.

[0037] Compared with the prior art, the application has the following advantages and beneficial effects:

[0038] The application realizes AI service process automation in a dynamic open environment through AI capability decoupling and hierarchical modeling, AI capability on-demand elastic combination optimization, and further promotes the efficient supply of AI capabilities by empowering digital employee platforms based on actual business needs; diversified AI capability decomposition and reconstruction are constructed, and an adaptive combination of AI capabilities is optimized to provide a systematic solution for AI applications in a dynamic open environment, thereby improving the adaptability and execution efficiency of AI technology in complex scenarios. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 Flowchart of the AI service process automatic generation method based on AI model combination optimization of the application

[0040] Figure 2 Overall architecture diagram of the application for constructing a multi-level AI capability decoupling and reconstruction module

[0041] Figure 3 Overall architecture diagram of the multi-level modeling of the application. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical scheme and advantages of the application clearer, the application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application and do not limit the application.

[0043] Example 1

[0044] As shown in the figure, the present embodiment provides an AI service process automatic generation method based on AI model combination optimization, comprising the following steps: Figure 1 S1: Construct a multi-level AI capability decoupling and reconstruction module, as shown in the figure, through hierarchical modeling from bottom to top and modular decoupling from top to bottom, realize dynamic mapping of AI atomic capability layer, AI modular production capability layer, AI general capability layer and AI application business layer, support dynamic decoupling in complex business scenarios;

[0045] Figure 2 As shown in the figure, the multi-level modeling specifically includes:

[0046] As shown in the figure, the multi-level modeling specifically includes: Figure 3 ​​

[0047] (1) AI atomic ability layer, used for refining business requirements and realizing atomic-level optimization of AI capabilities, including OCR, NLP, visual recognition, and intelligent optimization basic modules;

[0048] (2) AI modular production ability layer, used for generating modular AI capability components, supporting dynamic integration of multi-source heterogeneous data, and generating reusable AI components through feature cross, model fusion, and automatic machine learning technology;

[0049] (3) AI modular general ability layer, dynamically assembling AI components based on business requirements, forming computer vision, natural language processing function units, constructing general AI capability modules, and supporting functional combination of business scenarios;

[0050] (4) AI application business layer, mapping functional units to actual business processes through super heuristic algorithms, generating multi-AI process application combinations according to actual business requirements, and meeting multi-scenario business requirements;

[0051] In this embodiment, for complex business scenarios, a dynamic AI capability decoupling mode with strong business direction is constructed based on the dynamic decoupling mechanism of AI business capabilities, to realize efficient matching of AI capabilities and business requirements; by constructing a functional integrated multi-AI application capability decoupling system, combined with intelligent AI capability combination algorithms, a comprehensive intelligent AI application capability attribution and decision system is built, providing systematic support for AI capability invocation in complex business scenarios; in addition, based on the programmable AI general capability combination decoupling, an AI capability combination component with high reusability is constructed, ultimately forming a multi-level general AI capability decoupling framework with intelligent and flexible characteristics, which significantly improves the reusability, expandability, and business adaptability of AI capabilities, providing strong technical support for AI applications in a dynamic and open environment;

[0052] In this embodiment, a non-coupled module-level AI capability component is constructed based on the module-level granularity of the basic AI capability supply topology, aiming to realize efficient supply and flexible invocation of AI capabilities, and providing modular support for complex business scenarios; at the same time, based on the general componentized AI application combination and business deployment, an adaptive multi-AI service capability and business deployment process is constructed, significantly improving the adaptability and deployment efficiency of AI applications in a dynamic environment; in addition, a multi-level AI reconstruction model performance evaluation system is constructed through the modular evaluation mechanism of the multi-level AI reconstruction model, comprehensively optimizing the response speed and collaboration capability in dynamic AI service scenarios, providing a systematic solution for modular design, intelligent deployment, and dynamic optimization of AI capabilities.

[0053] In this embodiment, AI capabilities such as OCR, NLP, computer vision, etc. are decoupled into independent functional modules to enable flexible combination and reuse. This modular design not only reduces development and deployment costs, but also improves system scalability and adaptability, for example: when processing merged text, the OCR module can be called to convert scanned contract pictures into text data, and then the NLP module can be used for content analysis, such as extracting contract amount, date, and other key information. When identity verification is required, the OCR module can be used to read user ID card information, and the computer vision module can be used for face comparison to complete identity verification. Through this modular design, AI capabilities can be flexibly called according to actual needs, avoiding repeated development and significantly improving development efficiency.

[0054] In this embodiment, in order to cope with complex environments with multiple scenarios and multiple needs, based on super-heuristic algorithms such as genetic algorithms and particle swarm optimization, dynamic integration and combination optimization of AI capabilities are realized. Specifically, this method dynamically adjusts three dimensions of data, features, and models to ensure the efficiency and adaptability of AI capabilities, for example: in a recommendation system that needs to dynamically optimize the recommendation model based on user behavior data. Through super-heuristic algorithms, the optimal combination can be selected from multiple recommendation models such as collaborative filtering and deep learning models to meet the personalized needs of users while improving recommendation accuracy. In industrial scenarios, fault detection systems need to analyze multiple sensor data in real time. By dynamically combining different AI capabilities such as time series analysis and anomaly detection models, the fault source can be quickly located and the detection accuracy can be optimized. This optimization method not only improves the adaptability of AI capabilities, but also dynamically adjusts resource allocation according to scenario needs to ensure efficient operation of the system in complex environments.

[0055] S2: Build an elastic AI capability combination optimization module, based on quality of service requirements, propose an AI capability combination optimization method for three dimensions of data, features, and models, and through proxy model driven evolutionary computation optimization, realize the reuse, combination and arrangement of AI capabilities, including:

[0056] S21: Build a multi-dimensional QoS evaluation proxy model by superimposing RBF kernel function and linear kernel function;

[0057] S22: Use evolutionary computation methods to generate candidate model populations, and update the model library by combining proxy evaluation and real evaluation hybrid mechanisms;

[0058] In this embodiment, based on the quality of service requirements, by constructing an AI capability combination optimization framework facing data, features and models, efficient matching and dynamic adjustment of AI capability and business requirements are realized. The elastic AI capability combination optimization method optimizes data preprocessing and feature extraction from the data level, improves the accuracy of feature selection and fusion from the feature level, and enhances the model performance and adaptability from the model level, thereby comprehensively improving the flexibility and service quality of AI capability combination. At the same time, aiming at the challenge of difficult evaluation and optimization of AI capability combination, an evolutionary computation combination optimization method driven by a proxy model is proposed. This method realizes the rapid evaluation and optimization of complex AI capability combination by constructing a proxy model and combining evolutionary computation algorithm, which significantly reduces the calculation cost and time overhead. On this basis, flexible AI capability reuse, combination and arrangement are further realized, providing an efficient and flexible solution for AI application in a dynamic and open environment.

[0059] S3: Construct an AI business process automation generation module; based on data-driven process mining and super heuristic algorithm, generate and optimize AI service containerization deployment scheme, support AI business process automation generation in a dynamic and open environment, specifically including:

[0060] S31: Extract high-order features of historical business processes through process mining technology;

[0061] S32: Learn super heuristic rules using genetic programming method to dynamically generate containerization deployment scheme;

[0062] In this embodiment, the data-driven super heuristic AI application process automatic generation and deployment method aims to solve the intelligent construction and optimization of AI application process in multiple dynamic scenarios. This method extracts key process patterns from massive business data through data-driven technology combined with process mining algorithm, and realizes the automatic generation and optimization of the process based on super heuristic algorithm. Specifically, first, data-driven technology is used to deeply analyze business data in multiple dynamic scenarios to identify key nodes and dependency relationships in the process. Second, the process is intelligently optimized through a super heuristic algorithm to generate an efficient and flexible AI application process. Finally, combined with service containerization technology, the process is quickly deployed and dynamically adjusted, significantly improving the adaptability and execution efficiency of AI application in complex environments.

[0063] In this embodiment, by constructing a service containerization deployment framework and dynamically optimizing the deployment scheme with a super heuristic algorithm, the automatic generation and efficient deployment of AI business processes in a dynamically open complex environment are realized. The super heuristic algorithm is used to optimize the service containerization deployment scheme in multiple objectives, including resource utilization, deployment efficiency, and system stability, so as to realize the rapid generation and deployment of AI business processes in a complex environment, and provide a systematic solution for the construction and deployment of AI application processes in multiple dynamic scenarios.

[0064] In this embodiment, through data-driven technology and process mining, the automatic generation and intelligent deployment of business processes in a dynamically open environment are realized. Specifically, first, the massive business data is deeply analyzed to extract key process patterns, and then the process is optimized by combining a super heuristic algorithm, and finally an efficient AI application process is generated and quickly deployed. For example, the customer service system can analyze the customer's historical consultation data to mine the processing flow of common problems (such as account inquiry, password reset, etc.), and then use process automation technology to generate the corresponding AI customer service process and quickly go online through containerization deployment to realize the instant response to customer demand. In the logistics scenario, the optimal logistics scheduling process can be automatically generated by analyzing order data and transportation paths, and combined with containerization technology to dynamically adjust the transportation scheme, significantly improving logistics efficiency. This process automatic generation and deployment method not only improves the business response speed, but also can be adjusted in real time according to environmental changes to ensure the flexibility and stability of the system in a dynamically open environment.

[0065] S4: Integration on the distributed digital employee platform, verify the performance of the integrated system in a dynamically open environment, realize on-demand supply of AI capabilities and adaptive business orchestration.

[0066] In this embodiment, through multi-scenario testing, it performs well in accuracy, robustness, security, and response speed, can accurately match business needs, dynamically generate efficient AI application processes, and run stably in complex environments, while ensuring the security of process generation and deployment through multi-level security mechanisms, providing an efficient, flexible, and reliable solution for AI applications in multiple dynamic scenarios.

[0067] Embodiment 2

[0068] This embodiment provides an AI business process automation generation system based on AI model combination optimization, which is used to realize the AI business process automation generation method based on AI model combination optimization of the above-mentioned embodiment 1, including: a multi-level AI capability decoupling and reconstruction module, an elastic AI capability combination optimization module, and an AI business process automation generation module.

[0069] In this embodiment, the multi-level AI capability decoupling and reconstruction module realizes dynamic mapping of the AI atomic capability layer, the modular production capability layer, the general capability layer and the application business layer through bottom-up hierarchical modeling and top-down modular decoupling.

[0070] In this embodiment, the elastic AI capability combination optimization module proposes an AI capability combination optimization method based on three dimensions of data, features and models based on service quality requirements, and realizes reuse, combination and arrangement of AI capabilities through proxy model driven evolutionary calculation optimization.

[0071] Specifically, the elastic AI capability combination optimization module includes a proxy model driven evolutionary calculation optimization unit, a QoS evaluation unit, and a model library management unit.

[0072] The proxy model driven evolutionary calculation optimization unit is used to realize rapid evaluation and optimization of AI capability combination with the aid of a proxy model.

[0073] The QoS evaluation unit is used to evaluate the performance of AI capability combination based on a data driven multi-dimensional service quality evaluation proxy model.

[0074] The model library management unit is used to maintain the AI model library and support dynamic updating and combination of candidate models.

[0075] In this embodiment, the AI business process automation generation module generates and optimizes AI service container deployment schemes based on data driven process mining and super heuristic algorithms, and supports AI business process automation generation in a dynamically open environment.

[0076] Specifically, the AI business process automation generation module includes a process mining unit, a process optimization unit, and a containerization deployment unit.

[0077] The process mining unit is used to transparentize and twin the business process based on log data, operation data and basic data.

[0078] The process optimization unit is used to mine business logic and generate process optimization rules based on super heuristic algorithms, specifically including:

[0079] (1) Construct a reference process model pool based on process twinning technology, and design an adaptability function combined with QoS metrics.

[0080] (2) Encapsulate AI services using containerization technology, and dynamically adjust the deployment strategy according to real-time resource status.

[0081] The containerization deployment unit is used to modularize and efficiently deploy the optimized process.

[0082] Embodiment 3

[0083] The embodiment provides a storage medium, which can be a ROM, a RAM, a magnetic disk, an optical disk or the like storage medium, and the storage medium stores one or more programs, and the programs are executed by a processor to implement the AI service process automatic generation method based on AI model combination optimization in the embodiment 1.

[0084] Embodiment 4

[0085] The embodiment provides a computing device, which can be a desktop computer, a notebook computer, a smart phone, a PDA handheld terminal, a tablet computer or other terminal device with a display function, and the computing device comprises a processor and a memory, the memory stores one or more programs, and the processor executes the programs stored in the memory to implement the AI service process automatic generation method based on AI model combination optimization in the embodiment 1.

[0086] The above embodiment is a preferred embodiment of the present application, but the embodiment of the present application is not limited to the above embodiment, and any change, modification, replacement, combination, simplification made without departing from the spirit and principle of the present application should be an equivalent replacement mode, and all are included in the protection scope of the present application.

Claims

1. An AI business process automation generation method based on AI model combination optimization, characterized in that, Comprise the following steps: A multi-level AI capability decoupling and reconstruction module is constructed to realize dynamic mapping of an AI atomic capability layer, an AI modular production capability layer, an AI general capability layer, and an AI application business layer through bottom-up hierarchical modeling and top-down modular decoupling; An elastic AI capability combination optimization module is constructed to perform AI capability combination optimization in three dimensions of data, features, and models based on service quality requirements, to realize reuse, combination, and arrangement of AI capabilities through proxy model driven evolutionary computation optimization; An AI business process automation generation module is constructed to generate and optimize AI service containerization deployment schemes based on data-driven process mining and super heuristic algorithms, to realize AI business process automation generation. 2.The AI business process automation generation method based on AI model combination optimization according to claim 1, wherein, The AI atomic capability layer is configured to realize atomic-level optimization of AI capabilities. The AI modular production capability layer is configured to generate modular AI capability components and reusable AI components. The AI general capability layer is configured to assemble AI components based on business requirements to form functional units. The AI application business layer is configured to map the functional units to business processes through super heuristic algorithms. 3.The AI business process automation generation method based on AI model combination optimization according to claim 1, wherein, Through proxy model driven evolutionary computation optimization, reuse, combination, and arrangement of AI capabilities are realized, specifically including: A multi-dimensional QoS evaluation proxy model is constructed through superposition of RBF kernel functions and linear kernel functions to evaluate the performance of AI capability combination in QoS; An evolutionary computation method is used to generate a candidate model population, and a hybrid mechanism of proxy evaluation and real evaluation is used to update the model library to complete the update and combination of the candidate model. 4.The AI business process automation generation method based on AI model combination optimization according to claim 1, wherein, Based on data-driven process mining and super heuristic algorithms, AI service containerization deployment schemes are generated and optimized, specifically including: High-order features of historical business processes are extracted through process mining technology; A genetic programming method is used to learn super heuristic rules to generate containerization deployment schemes.

5. An AI business process automation generation system based on AI model combination optimization, characterized by, The AI business process automation generation method based on AI model combination optimization of any one of claims 1-4 comprises a multi-level AI capability decoupling and reconstruction module, an elastic AI capability combination optimization module, and an AI business process automation generation module. The multi-level AI capability decoupling and reconstruction module is configured to realize dynamic mapping of an AI atomic capability layer, an AI modular production capability layer, an AI general capability layer, and an AI application business layer through bottom-up hierarchical modeling and top-down modular decoupling. The elastic AI capability combination optimization module is configured to perform AI capability combination optimization in three dimensions of data, features, and models based on service quality requirements, to realize reuse, combination, and arrangement of AI capabilities through proxy model driven evolutionary computation optimization. The AI business process automation generation module is configured to generate and optimize AI service containerization deployment schemes based on data-driven process mining and super heuristic algorithms, to realize AI business process automation generation. 6.The AI business process automation generation system based on AI model combination optimization according to claim 5, wherein, The elastic AI capability combination optimization module comprises an evolutionary computation optimization unit, a QoS evaluation unit, and a model library management unit. The evolutionary computation optimization unit is configured to realize evaluation and optimization of AI capability combination through proxy model driving. The QoS evaluation unit is configured to evaluate the performance of the AI capability combination based on a proxy model. The model library management unit is configured to generate a candidate model population by using an evolutionary computing method, update the model library by combining a hybrid mechanism of proxy evaluation and real evaluation, and complete the updating and combination of the candidate model. 7.The AI business process automation generation system based on AI model combination optimization according to claim 5, wherein, The AI business process automation generation module includes a process mining unit, a process optimization unit, and a containerized deployment unit. The process mining unit is configured to perform business process transparency and twinning based on log data, operation data, and basic data. The process optimization unit is configured to mine business logic and generate process optimization rules based on a hyper-heuristic algorithm. The containerized deployment unit is configured to generate a containerized deployment scheme. 8.The AI business process automation generation system based on AI model combination optimization according to claim 7, wherein, The process optimization unit is configured to mine business logic and generate process optimization rules based on a hyper-heuristic algorithm, specifically including: Constructing a reference process model pool based on process twinning technology and designing an adaptive function based on QoS metrics; Encapsulating AI services based on containerization technology and adjusting deployment strategies according to real-time resource states.

9. A computer-readable storage medium storing a program, the program comprising instructions which, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 8. The program, when executed by the processor, implements the AI business process automation generation method based on AI model combination optimization as claimed in any one of claims 1-4.

10. A computer device comprising a processor and a memory for storing a processor executable program, characterized in that, The processor executes the program stored in the memory, and implements the AI business process automation generation method based on AI model combination optimization as claimed in any one of claims 1-4.

Images