Business process scheduling method and device, equipment and storage medium
Through the combination of deep learning model and LiteFlow framework, a business orchestration recommendation model is generated, which solves the problems of high complexity, low efficiency and poor adaptability of traditional business process scheduling and optimization technologies, and achieves efficient and flexible business process scheduling and optimization.
Patent Information
- Application Number
- CN202411500252.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-25
- Publication Date
- 2025-02-18
AI Technical Summary
Traditional business process scheduling and optimization technologies have problems such as high complexity, low efficiency, poor adaptability, difficulty in scaling and relying on manual tuning, and it is difficult to meet complex and dynamic business needs.
Using a combination of deep learning model and LiteFlow framework, we generate business orchestration recommendation models by collecting business data from the database, obtaining feature information, training deep learning models from the database and integrating LiteFlow framework, and automatically optimizing suggestions to reduce manual intervention.
It improves the execution efficiency and flexibility of business processes, reduces the dependence on manual intervention, and can make intelligent decisions and scheduling based on real-time business needs and system status, adapting to complex and dynamic business environments.
Smart Images
Figure CN119476912A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of software development, and in particular to a business process scheduling method, device, equipment and storage medium. Background Art
[0002] In the iterative process of software system development, various types of business processes will be generated to adapt to different business needs. For complex business processes, we need to simplify them, break down complex problems into small modules, and solve them one by one. The responsibilities of each module are relatively simple, which can ensure that future scalability and maintainability are relatively independent and simple.
[0003] Traditional business process scheduling and optimization technologies can achieve business splitting and decoupling through optimizing program design, reasonable design patterns, introducing rule engines, and expert system technologies, thereby simplifying business processes. However, this approach has the following disadvantages: 1. High complexity: Traditional methods often require a lot of manual participation and rule definition. When dealing with complex business process scheduling problems, it is difficult to cope with situations such as a large number of tasks and complex resource allocation.
[0004] 2. Low efficiency: Traditional methods are often inefficient when dealing with large-scale systems and complex business processes, and cannot achieve fast and intelligent scheduling and optimization.
[0005] 3. Poor adaptability: Traditional methods often lack adaptability and cannot make intelligent decisions and scheduling based on real-time business needs and system status.
[0006] 4. Difficult to expand: Traditional technologies are often difficult to expand and adapt to new business needs when faced with complex business processes and large-scale systems, which limits the flexibility and scalability of the system.
[0007] 5. Reliance on manual tuning: Some technologies based on rule engines or expert systems require manual tuning of rules and strategies, which consumes time and resources and cannot achieve automated optimization.
[0008] In summary, the technology of business process scheduling and optimization in traditional systems has some problems and challenges, and it is difficult to meet the increasingly complex and dynamic business needs. In order to solve the problems of low efficiency, poor adaptability, difficulty in expansion, and reliance on a large number of manual labor in traditional business process scheduling and optimization technology. Summary of the invention
[0009] Purpose of the invention: To propose a business process scheduling method, device, equipment and storage medium, which can improve the execution efficiency and flexibility of the process by combining a deep learning model with the LiteFlow framework, and reduce dependence on manual intervention by automatically generating optimization suggestions.
[0010] First, a business process scheduling method is proposed, and the steps are as follows: Collect business data samples from the database; Acquire characteristic information data from the service data sample, and convert the characteristic information data into a standard format containing service characteristic content; Selecting an open source deep learning network model; training the deep learning network model using the feature information data; Integrate the LiteFlow framework on the basis of the trained deep learning network model to generate a service orchestration recommendation model; The business orchestration recommendation model is connected to the business system, the business orchestration recommendation model recommends an optimal business process, and the business system deploys the optimal business process.
[0011] In a further embodiment of the first aspect, before acquiring the characteristic information data from the business data samples, it also includes data cleaning and normalization processing on the business data samples collected from the database to eliminate noise and outliers and ensure data quality.
[0012] Acquire characteristic information data from the business data sample that has undergone data cleaning and normalization processing, and convert the characteristic information data into a standard format containing business characteristic content; the standard format containing business characteristic content includes the following fields: Input, output, business execution time, and relationships between tasks.
[0013] In a further embodiment of the first aspect, the open source deep learning network model can be an artificial neural network ANN, a recurrent neural network RNN, a long short-term memory network LSTM, or a Transformer model.
[0014] When the open source deep learning network model selects a recurrent neural network RNN, the deep learning network model is trained using the feature information data, specifically including: Import the recurrent neural network RNN framework and related libraries; A model is constructed using a GRU layer and a Dropout layer, and overfitting is reduced by the Dropout layer; Select the predetermined loss function and optimizer, and select the evaluation indicator according to the actual needs of business process scheduling; Determine the number of samples to use in each training step and the number of times to iterate over the training dataset; Dividing the feature information data into a training set and a validation set according to a predetermined ratio; using the training set and the validation set as inputs of a recurrent neural network RNN, and starting to train the recurrent neural network RNN; After the recurrent neural network RNN training is completed, the recurrent neural network RNN is evaluated using a test data set to obtain the performance of the recurrent neural network RNN on unseen data.
[0015] In a further embodiment of the first aspect, the LiteFlow framework is integrated on the basis of the trained deep learning network model, specifically including: Build a project based on SpringBoot, introduce LiteFlow component dependencies, and edit the configuration file content; Create a LiteFlow component, use the LiteFlow component to define the process chain and process rules, and call the deep learning model to automatically infer the optimal process according to the context description of the process rule semantics; When the SpringBoot project configuration is completed, the process optimization is triggered through the API interface; the response data is provided according to the defined LiteFlow component output service, and the response data is recorded.
[0016] After the business orchestration recommendation model is connected to the business system, a monitoring tool is also set up to monitor the performance of the business system and the execution of the business process in real time; The performance of the business system and the execution status of the business process obtained in real time by the monitoring tool are regularly fed back to the business orchestration recommendation model; the business orchestration recommendation model performs performance evaluation on the business process and continuously collects incremental data; and the incremental data is used to train and adjust the business orchestration recommendation model.
[0017] According to a second aspect of the present invention, a business process scheduling device is proposed, which includes a database, a data acquisition component, a feature information acquisition component, a training unit, an integration component, and a deployment component. Historical business data is pre-stored in the database, and business data can be continuously supplemented during subsequent business process scheduling. The data acquisition component is used to collect business data samples from the database. The feature information acquisition component is used to obtain feature information data from the business data samples and convert the feature information data into a standard format containing business feature content. The training unit is based on an open source deep learning network model and uses feature information data to train the deep learning network model. The integration component is used to integrate the LiteFlow framework on the basis of the trained deep learning network model to generate a business orchestration recommendation model. The deployment component is used to connect the business orchestration recommendation model to the business system, and the business orchestration recommendation model recommends the optimal business process, and the business system deploys the optimal business process.
[0018] According to a third aspect of the present invention, an electronic device is provided, comprising: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the business process scheduling method as described in the first aspect is implemented.
[0019] According to a fourth aspect of the present invention, a computer-readable storage medium is provided, wherein at least one executable instruction is stored in the storage medium. When the executable instruction is executed on an electronic device, the electronic device executes the business process scheduling method as described in the first aspect.
[0020] Beneficial effects: The business process scheduling method proposed in the present invention identifies bottlenecks and inefficient links in business processes based on the prediction results of deep learning models. The process is dynamically adjusted according to the prediction results, business rules, and incremental business data. In the LiteFlow framework, scripts or configuration files are used to define business processes, including process nodes, conditional judgments, task execution, etc. The LiteFlow framework schedules the execution of each task according to the process definition, coordinates the interaction between various components, and connects with the capabilities of the large model. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 It is a flowchart of the business process scheduling method in the embodiment.
[0022] Figure 2 It is an initial flow chart of a data import business process in the embodiment.
[0023] Figure 3 It is one of the flowcharts of a certain data import business process after optimization and recommendation in the embodiment.
[0024] Figure 4 This is the second flowchart of a data import business process in the embodiment after optimization and recommendation.
[0025] Figure 5 It is a structural diagram of a business process scheduling device in an embodiment.
[0026] Figure 6 It is a schematic diagram of providing a business process scheduling method as a component with an API or as an independent system.
[0027] Figure 7 2 is a diagram showing the internal structure of an electronic device in an embodiment. DETAILED DESCRIPTION
[0028] In the following description, a large number of specific details are provided to provide a more thorough understanding of the present invention. However, it is apparent to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, in order to avoid confusion with the present invention, some technical features known in the art are not described.
[0029] The business process scheduling method proposed in the present invention is suitable for software systems that need to process complex business processes, especially in the fields of finance, supply chain management, intelligent manufacturing, etc., and can help enterprises improve the execution efficiency and effectiveness of business processes.
[0030] The technical points of the present invention mainly include: (1) Business process optimization and adjustment: Based on the prediction results of the deep learning model, identify bottlenecks and inefficient links in the business process. Dynamically adjust the process based on the prediction results, business rules, and incremental business data; (2) Business process definition and execution: Use scripts or configuration files in the LiteFlow framework to define business processes, including process nodes, conditional judgments, task execution, etc. The LiteFlow framework schedules the execution of each task based on the process definition, coordinates the interaction between various components, and connects with the capabilities of the large model.
[0031] The business process scheduling method is described below through multiple embodiments. Example 1
[0032] This embodiment discloses a complex business process optimization scheduling method using deep learning and LiteFlow. The method uses deep learning algorithms to analyze and optimize business processes. The overall business process diagram is shown in Figure 1 : Step 1: Data collection and preprocessing: The system process data is collected through the data collection and processing module and stored in the local database. The AI engine obtains the data samples from the local database for feature mining, obtains feature information, cleans, preprocesses, and formats the collected data to eliminate noise and outliers and ensure data quality. The data is feature extracted and converted to convert the original data into a format acceptable to the model, which may include input and output, business execution time, relationship between tasks and other feature contents.
[0033] Step 2: Model training based on deep learning: Based on the feature information data processed in step 1, a deep learning model is used for model training to generate a business orchestration recommendation model. The specific contents include: model construction, model compilation, and model training. The specific steps for each detail are as follows: (1) Model construction Import necessary libraries: Import deep learning frameworks and related libraries.
[0034] Define the model architecture: Use GRU layer and Dropout layer to build the model. The Dropout layer reduces overfitting and improves the generalization ability of the model.
[0035] (2) Model compilation Choose loss function and optimizer: Choose appropriate loss function and optimizer to improve the performance of the model.
[0036] Specify evaluation indicators: Select evaluation indicators based on task requirements.
[0037] (3) Model training Specify batch size and number of training epochs: Determine the number of samples to use for each training step and the number of times to iterate over the training dataset.
[0038] Call the model's fit method for training: take the training set and validation set as input and start model training.
[0039] Evaluate the model: After the model is trained, evaluate the model using the test dataset to get the performance of the model on unseen data.
[0040] Step 3: Integrate the deep learning model with the LiteFlow framework: Integrating the trained model with the LiteFlow framework requires combining the reasoning capabilities of the learning model with the process orchestration capabilities of the LiteFlow framework. The specific steps to implement the solution are: (1) Preparation: The trained model can be accessed; build a project based on the SpringBoot scaffolding to integrate the LiteFlow component; (2) Define the interaction between LiteFlow components and models: Create LiteFlow components, which are responsible for interacting with the model. Define process chains and process rules. Based on the context description of the semantics of process rules, the deep learning model's automatic reasoning and calculation capabilities can be used to recommend the optimal process. (3) Create a controller to trigger the process: After the SpringBoot project is configured, the process optimization is triggered through the API interface. According to the defined LiteFlow component output service, the response data is provided and the data is recorded to provide data support for the incremental training of the model.
[0041] Step 4: Business monitoring and feedback: In this step, some monitoring tools can be used to monitor the performance of the system and the execution of business processes in real time. The trained deep learning model can be used to evaluate the performance of the business process, and incremental data can be continuously collected for model training and adjustment to continuously improve the prediction accuracy of the model.
[0042] Step 5: System deployment and maintenance: After completing the above steps, this system has the ability to provide independent services. It also supports packaging this system into business components and using it in other projects. The system can be updated according to changes in requirements and continuously improved to adapt to new business requirements and changes. Example 2
[0043] Based on the above-mentioned embodiment 1, this embodiment takes Java language as an example to provide a detailed description and code examples for implementing the business process scheduling method.
[0044] Step 1: Data collection and processing: Collect real-time business data through the data collection and processing module, clean and extract features from the collected data, and implement data preprocessing logic, such as normalization and feature extraction. The data processing code example is as follows: @Component public class DataAcquisitionAndProcessor{ public List <processeddata>preprocessData(List <datapoint>rawData){ / / Implement data preprocessing logic, such as normalization, feature extraction, etc. return rawData.stream() .map(dataPoint ->new ProcessedData(dataPoint.getName(),normalize(dataPoint.getvalue()))) .collect(Collectors.toList()); } private double normalize(int value){ / / Simple normalization example return value 100.0; } } Step 2: Model training based on deep learning: Perform model training based on the feature information data processed in step 1 to generate a business orchestration recommendation model. At the same time, deploy the model so that it can provide computing power to the outside world. The deep training model code is as follows: @Component public class DeepLearningModel { private final RestTemplate restTemplate new RestTemplate(); private final String modelEndpoint "http: / / deep-learning-service / predict"; public PredictionResult predict(List <processeddata>inputData){ HttpHeaders headers = new HttpHeaders(); headers.setContentType(MediaType.APPLICATION_JSON); HttpEntity<List <processeddata>>requestEntity = new HttpEntity<>(inputData,headers); ResponseEntity <predictionresult>response =restTemplate.exchange(modelEndpoint, HttpMethod.POST,requestEntity,PredictionResult.class); return response.getBody(); } } Step 3: Integrate the deep learning model with the LiteFlow framework: (1) Build a project based on SpringBoot, complete the introduction of LiteFlow component dependencies, and improve the configuration file content; the project configuration code is as follows: # Configuration items in the SpringBoot configuration file liteflow.rule-source=classpath:flow.el The LiteFlow component configuration items are as follows: chain: THEN (optimizeProcess); (2) Define the LiteFlow process chain and process rules: Define the LiteFlow component to process business logic and deep learning model integration, and use this component to call the model's automatic reasoning and calculation capabilities. See the sample code below: @Component("optimizeProcess") public class OptimizeProcessComponent extends NodeComponent{ @Autowired private DataAcquisitionAndProcessor dataAcquisitionAndProcessor; @Autowired private DeepLearningModel deepLearningModel; @Override public void process()throws Exception { / / Data preprocessing List <processeddata>processedData = dataAcquisitionAndProcessor.preprocessData(rawData); / / Call the deep learning model for prediction PredictionResult prediction = deepLearningModel.predict(processedData); / / Adjust business processes based on pre-emptive measures if (prediction.requiresoptimization()){ adjustWorkflow(prediction); } / / Store the pre-processed results in the context for subsequent components to use this.getSlot().setRequestData("predictionResult",prediction); } private void adjustWorkflow(PredictionResult prediction){ / / Implement business process optimization logic, such as weekly task priorities, update allocation resources, etc. / / do somethings … } } (3) Create a controller to trigger the process: provide an API interface for other services to call. See the sample code below: @RestController public class ProcessController { @Autowired private FlowExecutor flowExecutor; @GetMapping(" / optimizeProcess") public String optimizeProcess(){ flowExecutor.execute2Resp("chain","arg"); return "Process optimized"; } } Step 4: After calculation and recommendation by this system, the business process can be optimized. This article takes a data import business process as an example. The initial established business process is shown in Figure 2 ; The business process optimized and recommended by this system can be found in Figure 3 or Figure 4 .
[0045] Figure 7 The internal structure diagram of the electronic device. The electronic device includes a processor, a memory, a communication interface and a communication bus. The processor, the memory and the communication interface communicate with each other through the communication bus. The memory is used to store at least one executable instruction, which enables the processor to execute Figure 1 The running steps of the business process scheduling method.
[0046] The electronic device may also communicate with one or more external devices (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may communicate with one or more devices that enable a user to interact with the electronic device, and / or may communicate with any device that enables the electronic device to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication may be performed through an input / output (I / O) interface. In addition, the electronic device may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter. The network adapter communicates with other modules of the electronic device through a bus. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.
[0047] This embodiment also provides a computer-readable storage medium, in which at least one executable instruction is stored, and when the executable instruction is executed on an electronic device, the electronic device performs the operation of the database optimization method described in the above embodiment. More specific examples of computer-readable storage media in the present disclosure may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0048] Computer readable storage media may include data signals propagated in baseband or as part of a carrier wave, which carry readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than readable storage media, which may send, propagate, or transmit programs for use by or in conjunction with an instruction execution system, apparatus, or device. Example 3
[0049] See Figure 5 As shown, this embodiment discloses a business process scheduling device 600, which includes a database 601, a data acquisition component 602, a feature information acquisition component 603, a training unit 604, an integration component 605, and a deployment component 606. The database 601 is pre-stored with historical business data, and the business data can be continuously supplemented during the subsequent business process scheduling. The data acquisition component 602 is used to collect business data samples from the database 601. The feature information acquisition component 603 is used to obtain feature information data from the business data sample and convert the feature information data into a standard format containing business feature content. The training unit 604 is based on an open source deep learning network model and uses feature information data to train the deep learning network model. The integration component 605 is used to integrate the LiteFlow framework on the basis of the trained deep learning network model to generate a business orchestration recommendation model. The deployment component 606 is used to connect the business orchestration recommendation model to the business system, and the business orchestration recommendation model recommends the optimal business process, and the business system deploys the optimal business process.
[0050] Figure 6 The diagram shows the business process scheduling method as a component that provides an API or an independent system, which includes the operating environment, storage layer, business layer, control layer, and display layer. The storage layer is a Mysql database, and the business layer includes a data acquisition processing unit, LiteFlow process chain and process rules, and a data model service. The control layer includes SpringBoot services.
[0051] In summary, the present invention provides the following technical effects: (1) Cost reduction and efficiency improvement: Improve system response speed and resource utilization, reduce operating costs, and improve business efficiency; (2) Technological innovation: Integrate AI technology with traditional frameworks to improve the stable output capabilities of traditional frameworks; (3) Reducing the pressure on coders: Compared with previous implementations, this invention uses the power of AI to output qualified content, which developers can fine-tune or use directly.
[0052] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification. Although the present invention has been represented and described with reference to specific preferred embodiments, it shall not be interpreted as a limitation of the present invention itself. Various changes can be made to it in form and detail without departing from the spirit and scope of the present invention defined in the attached claims.< / processeddata> < / predictionresult> < / processeddata> < / processeddata> < / datapoint> < / processeddata>
Claims
1. A business process scheduling method, characterized in that: The steps include: Collect business data samples from the database; Acquire characteristic information data from the service data sample, and convert the characteristic information data into a standard format containing service characteristic content; Selecting an open source deep learning network model; training the deep learning network model using the feature information data; Integrate the LiteFlow framework on the basis of the trained deep learning network model to generate a service orchestration recommendation model; The business orchestration recommendation model is connected to the business system, the business orchestration recommendation model recommends an optimal business process, and the business system deploys the optimal business process.
2. A business process scheduling method according to claim 1, characterized in that: Before acquiring the characteristic information data from the service data sample, the method further includes: Perform data cleaning and normalization on the business data samples collected from the database.
3. A business process scheduling method according to claim 2, characterized in that: Acquire characteristic information data from the business data sample that has undergone data cleaning and normalization processing, and convert the characteristic information data into a standard format containing business characteristic content; the standard format containing business characteristic content includes the following fields: Input, output, business execution time, and relationships between tasks.
4. A business process scheduling method according to claim 1, characterized in that: The open source deep learning network model can be an artificial neural network ANN, a recurrent neural network RNN, a long short-term memory network LSTM, or a Transformer model.
5. A business process scheduling method according to claim 4, characterized in that: When the open source deep learning network model selects a recurrent neural network RNN, the deep learning network model is trained using the feature information data, specifically including: Import the recurrent neural network RNN framework and related libraries; A model is constructed using a GRU layer and a Dropout layer, and overfitting is reduced by the Dropout layer; Select the predetermined loss function and optimizer, and select the evaluation indicator according to the actual needs of business process scheduling; Determine the number of samples to use in each training step and the number of times to iterate over the training dataset; Dividing the feature information data into a training set and a validation set according to a predetermined ratio; using the training set and the validation set as inputs of a recurrent neural network RNN, and starting to train the recurrent neural network RNN; After the recurrent neural network RNN training is completed, the recurrent neural network RNN is evaluated using a test data set to obtain the performance of the recurrent neural network RNN on unseen data.
6. A business process scheduling method according to claim 1, characterized in that: The LiteFlow framework is integrated on the basis of the trained deep learning network model, specifically including: Build a project based on SpringBoot, introduce LiteFlow component dependencies, and edit the configuration file content; Create a LiteFlow component, use the LiteFlow component to define the process chain and process rules, and call the deep learning model to automatically infer the optimal process according to the context description of the process rule semantics; When the SpringBoot project configuration is completed, the process optimization is triggered through the API interface; the response data is provided according to the defined LiteFlow component output service, and the response data is recorded.
7. A business process scheduling method according to claim 5, characterized in that: After the service arrangement recommendation model is connected to the service system, the following steps are also included: Set up monitoring tools to monitor the performance of business systems and the execution of business processes in real time; Regularly feeding back the performance of the business system and the execution status of the business process obtained in real time by the monitoring tool to the business orchestration recommendation model; The business orchestration recommendation model performs performance evaluation on the business process and continuously collects incremental data; The incremental data is used to train and adjust the service orchestration recommendation model.
8. A business process scheduling device, characterized in that: include: Database; the database stores historical business data in advance, and can continuously supplement business data during subsequent business process scheduling; Data acquisition component, used to collect business data samples from the database; A feature information acquisition component, used to acquire feature information data from the service data sample and convert the feature information data into a standard format containing service feature content; A training unit, based on an open source deep learning network model, uses the feature information data to train the deep learning network model; Integration component, used to integrate LiteFlow framework on the basis of trained deep learning network model to generate business orchestration recommendation model; The deployment component is used to connect the business orchestration recommendation model to the business system, and the business orchestration recommendation model recommends the optimal business process, and the business system deploys the optimal business process.
9. An electronic device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; when the processor executes the computer program instructions, the business process scheduling method according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that: The storage medium stores at least one executable instruction, and when the executable instruction is executed on the electronic device, the electronic device executes the business process scheduling method according to any one of claims 1 to 7.