RPA process optimization method based on large language model

Abstract

The invention discloses an RPA process optimization method based on a large language model. The method comprises the following steps: S1, carrying out preprocessing and structuring on an existing RPA process; s2, mapping the existing structured RPA flow into a knowledge graph; s3, analyzing the mapped knowledge graph by using LLM (Logistics Language Model); s4, adjusting and optimizing the existing RPA process according to the analysis result in the step S3 to obtain an optimized RPA process; and S5, re-converting the optimized RPA flow into a script or a command sequence which can be executed by the RPA, and re-deploying the optimized RPA flow on the RPA platform by adopting an RPA platform development tool. According to the method, the efficiency of the optimized RPA process in the actual application process can be effectively improved, the operation and maintenance cost is reduced, and the method has great practical value.

Description

Technical Field

[0001] The present invention relates to the technical field of Robotic Process Automation (RPA), and in particular to an RPA process optimization method based on a large language model. Background Art

[0002] With the continuous development of technology and the rapid increase in data size, traditional business process management methods face new challenges, especially in environments with complex processing requirements and unstructured information. In order to improve efficiency, RPA technology has been widely adopted, which significantly improves the efficiency of business operations by simulating human behavior to automatically perform repetitive tasks. However, traditional RPA systems mainly rely on preset rules and lack the necessary adaptability and flexibility, so they perform poorly when faced with tasks that require deep decision-making and expertise.

[0003] To solve this problem, knowledge graph technology was introduced to enhance the decision support capabilities of RPA. By creating a network diagram of entities and their relationships, knowledge graphs help machines understand complex business knowledge and support more sophisticated decision-making processes.

[0004] Despite this, existing technologies still have many limitations when combining knowledge graphs with RPA. On the one hand, many knowledge graphs are static, not updated in a timely manner, and difficult to adapt to the rapidly changing business environment; on the other hand, most of these combined systems focus on information retrieval and are not sufficient to support complex decision-making needs, which limits the performance of RPA in performing difficult tasks. In addition, existing technologies that fail to optimize data processing and response speed also affect the real-time and efficiency of operations.

[0005] Therefore, it is an urgent problem to design a new RPA process optimization method in combination with the Large Language Model (LLM) to overcome the above defects.

[0006] It should be understood that the above statements merely provide background technology related to the present invention and do not necessarily constitute prior art. Summary of the invention

[0007] The purpose of the present invention is to provide an RPA process optimization method based on a large language model, which integrates the advanced analysis capabilities of the large language model and the execution efficiency of the RPA process, thereby improving the intelligence level of RPA process automation.

[0008] In order to achieve the above object, the present invention provides an RPA process optimization method based on a large language model, which specifically includes the following steps:

[0009] S1. Preprocess and structure the existing RPA process;

[0010] S2. Map the structured existing RPA process into a knowledge graph;

[0011] S3. Use LLM to analyze the mapped knowledge graph;

[0012] S4. Adjust and optimize the existing RPA process according to the analysis results in S3 to obtain an optimized RPA process;

[0013] S5. Convert the optimized RPA process back into an RPA executable script or command sequence, and redeploy the optimized RPA process on the RPA platform using the RPA platform development tool.

[0014] Preferably, in S1, the structuring is specifically manifested as: identifying each step and decision point in the existing RPA process and converting them into a standardized format.

[0015] Preferably, the standardized format is a data structure in JSON or XML format.

[0016] Among them, if the existing RPA process involves business rules and logical relationships, the standardized format is described in pseudocode or flowchart.

[0017] Preferably, in S2, the nodes in the knowledge graph represent the steps and decision points in the existing RPA process; the edges in the knowledge graph represent the dependencies and data flows between the nodes.

[0018] Preferably, the steps in the existing RPA process specifically include: data input, adopted processing functions and output results.

[0019] Preferably, the knowledge graph is mapped using a knowledge graph tool, and the knowledge graph tool is Neo4j.

[0020] Preferably, in S3, the LLM is able to identify logical redundancies and potential optimization points in the existing RPA process and make corresponding modification suggestions.

[0021] Preferably, in S4, adjusting and optimizing the existing RPA process includes: adjusting the order of steps in the existing RPA process, merging redundant operations in the existing RPA process, and simplifying complex conditional logic in the existing RPA process.

[0022] Preferably, the optimized RPA process can be tested through simulation tools to ensure that the optimized RPA process achieves efficiency improvement.

[0023] Preferably, in S5, the RPA platform development tool is UiPath or Automation Anywhere.

[0024] In summary, compared with the prior art, the RPA process optimization method based on a large language model provided by the present invention has at least the following beneficial effects:

[0025] (1) The present invention builds a system that can use LLM to understand and adapt to complex business environments, ensuring that the automated process can efficiently handle emergencies and business changes, reducing dependence on manual intervention and having strong adaptability;

[0026] (2) The present invention utilizes the natural language processing capabilities of LLM to improve the efficiency and accuracy of the RPA process in parsing and processing unstructured data, thereby expanding its scope and depth of application;

[0027] (3) The present invention develops a dynamic learning and process update mechanism, which enables the RPA system to automatically adjust the process according to business development and data feedback, reducing maintenance costs and improving response speed;

[0028] (4) The present invention uses a large language model to intelligently review and optimize the RPA process mapped in the knowledge graph, identify redundant and inefficient links in the process, and put forward specific optimization suggestions, providing a new and automated process optimization method, making process optimization more efficient and intelligent;

[0029] (5) The RPA process optimized by the present invention is more reasonable, reducing the possibility of human errors and system failures, making the RPA process more stable and reliable during execution. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 This is a specific flow chart of the RPA process optimization method based on a large language model of the present invention;

[0031] Figure 2 It is the existing RPA process preprocessing and structured flow chart in the present invention;

[0032] Figure 3 A structural diagram that maps the existing RPA process in the present invention to a knowledge graph;

[0033] Figure 4 This is a schematic diagram of the RPA process analysis based on LLM in the present invention;

[0034] Figure 5 A flowchart for adjusting and optimizing the existing RPA process in the present invention;

[0035] Figure 6 This is a schematic diagram of the output of the optimized RPA process in the present invention. DETAILED DESCRIPTION

[0036] The following is combined with Figure 1 ~Attached Figure 6 , the present invention is further explained by describing a preferred specific embodiment in detail.

[0037] It should be noted that the drawings are in a very simplified form and use non-precise proportions. They are only used to conveniently and clearly assist in explaining the embodiments of the present invention, and are not used to limit the conditions for the implementation of the present invention. Therefore, they have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical content disclosed by the present invention.

[0038] It should be noted that, in the present invention, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only the elements explicitly listed, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0039] like Figure 1 As shown, the present invention provides an RPA process optimization method based on a large language model, which specifically includes the following steps:

[0040] S1. Preprocess and structure the existing RPA process;

[0041] S2. Map the structured existing RPA process into a knowledge graph;

[0042] S3. Use LLM to analyze the mapped knowledge graph;

[0043] S4. Adjust and optimize the existing RPA process according to the analysis results in S3 to obtain an optimized RPA process;

[0044] S5. Convert the optimized RPA process back into an RPA executable script or command sequence, and redeploy the optimized RPA process on the RPA platform using the RPA platform development tool.

[0045] Further, such as Figure 2As shown, in S1, the existing RPA process is structured, and its specific contents include: identifying each step and decision point in the existing RPA process and converting them into a standardized format; wherein the standardized format is generally a data structure in JSON or XML format; however, it should be noted that for the business rules and logical relationships involved in the existing RPA process, the standardized format is described in detail in the form of pseudocode or flowchart to ensure that the business rules and logical relationships involved are accurately and unambiguously recorded.

[0046] Further, such as Figure 3 As shown, in S2, the nodes in the knowledge graph can represent the various steps and decision points in the existing RPA process, and the edges in the knowledge graph can represent the logical relationship and data flow between the nodes; wherein the steps specifically include: data input, adopted processing functions and output results.

[0047] It is understandable that a knowledge graph tool can be used to create and store a knowledge graph to ensure that each part of the existing RPA process can be clearly mapped. In a preferred embodiment of the present invention, the knowledge graph tool uses Neo4j.

[0048] Further, such as Figure 4 As shown, in S3, LLM analysis is performed on the mapped knowledge graph to identify logical redundancies and potential optimization points in the existing RPA process, and corresponding modification suggestions are proposed.

[0049] It can be understood that the purpose of the preprocessing and structuring is to enable LLM to effectively understand and analyze the existing RPA process.

[0050] Specifically, the execution method of the LLM is to make specific queries to the LLM, in other words, to ask the LLM questions such as "Which steps can be merged?" or "Where can the conditional judgment be simplified?", so that the LLM will analyze the business rules and logical relationships of the existing RPA process and make specific improvement suggestions. It is understandable that the aforementioned improvement suggestions include specific operations such as structural adjustment of the existing RPA process and reconfiguration of resources.

[0051] Further, such as Figure 5As shown, in S4, according to the improvement suggestions provided by the LLM, the order of steps in the existing RPA process is adjusted, the redundant operations in the existing RPA process are merged, and the complex conditional logic in the existing RPA process is simplified to obtain the optimized RPA process. It should be noted that the optimized RPA process can be tested using simulation tools to ensure that each improvement can bring the expected efficiency improvement. At the same time, the impact of each RPA process improvement on the previous RPA process should also be recorded to evaluate the effect of the optimization and serve as a basis for subsequent further optimization.

[0052] Further, such as Figure 6 In S5, first, the optimized RPA process needs to be converted back into an RPA executable script or command sequence to ensure that all RPA process optimization improvements are accurately converted into a format that the RPA platform can understand and execute; secondly, the RPA platform development tool is used to redeploy and test the optimized RPA process, and put it into normal actual use after the test is completed.

[0053] Among them, in a preferred embodiment of the present invention, the RPA platform development tool adopts UiPath or AutomationAnywhere.

[0054] In summary, the present invention provides an RPA process optimization method based on a large language model, which preprocesses and structures the existing RPA process, converts it into a knowledge graph, and then uses the large language model to analyze and optimize it, and dynamically adjusts and updates it to achieve intelligent RPA process optimization, which significantly improves the adaptability and intelligent decision-making ability of the optimized RPA process, and can effectively improve the efficiency of the optimized RPA process in actual application and reduce operation and maintenance costs.

[0055] Although the content of the present invention has been described in detail through the above preferred embodiments, it should be appreciated that the above description should not be considered as a limitation of the present invention. After reading the above content, it will be apparent to those skilled in the art that various modifications and substitutions of the present invention will occur. Therefore, the protection scope of the present invention should be limited by the appended claims.

Claims

1. A RPA process optimization method based on a large language model, characterized in that: The specific steps include: S1. Preprocess and structure the existing RPA process; S2. Map the structured existing RPA process into a knowledge graph; S3. Use LLM to analyze the mapped knowledge graph; S4. Adjust and optimize the existing RPA process according to the analysis results in S3 to obtain an optimized RPA process; S5. Convert the optimized RPA process back into an RPA executable script or command sequence, and redeploy the optimized RPA process on the RPA platform using the RPA platform development tool.

2. The RPA process optimization method based on a large language model according to claim 1, characterized in that: In S1, the structuring is specifically manifested as: identifying each step and decision point in the existing RPA process and converting them into a standardized format.

3. The RPA process optimization method based on a large language model as described in claim 2 is characterized in that: The standardized format is a data structure in JSON or XML format; Among them, if the existing RPA process involves business rules and logical relationships, the standardized format is described in pseudocode or flowchart.

4. The RPA process optimization method based on a large language model as described in claim 2 is characterized in that: In S2, the nodes in the knowledge graph represent steps and decision points in the existing RPA process; The edges in the knowledge graph represent the dependencies and data flows between the nodes.

5. The RPA process optimization method based on a large language model as claimed in claim 4, characterized in that: The steps in the existing RPA process specifically include: data input, adopted processing functions and output results.

6. The RPA process optimization method based on a large language model as claimed in claim 4, characterized in that: The knowledge graph is mapped using a knowledge graph tool, and the knowledge graph tool is Neo4j.

7. The RPA process optimization method based on a large language model according to claim 1, characterized in that: In S3, the LLM can identify logical redundancies and potential optimization points in the existing RPA process and make corresponding modification suggestions.

8. The RPA process optimization method based on a large language model according to claim 1, characterized in that: In the above-mentioned S4, the content of adjusting and optimizing the existing RPA process includes: adjusting the step sequence of the existing RPA process, merging redundant operations in the existing RPA process, and simplifying the complex conditional logic in the existing RPA process.

9. The RPA process optimization method based on a large language model as claimed in claim 8, characterized in that: The optimized RPA process can be tested through simulation tools to ensure that the optimized RPA process achieves efficiency improvement.

10. The RPA process optimization method based on a large language model according to claim 1, characterized in that: In S5, the RPA platform development tool is UiPath or AutomationAnywhere.

Images