CN-121979802-A - Test case generation and intelligent management system based on multi-agent cooperation

CN121979802ACN 121979802 ACN121979802 ACN 121979802ACN-121979802-A

Abstract

The invention belongs to the technical field of artificial intelligence, and particularly relates to a test case generation and intelligent management system based on multi-agent cooperation. The system comprises an input and configuration module, an agent cooperation module and an RPA robot execution module, wherein the input and configuration module is used for interacting with a user and receiving a requirement document uploaded by the user and configuration information related to test case management, the agent cooperation module is used for generating and evaluating test cases meeting quality requirements through cooperation of a plurality of agents based on the requirement document and the configuration information, and the RPA robot execution module is used for triggering automatic warehousing operation of the test cases after the test cases are evaluated and passed and simultaneously importing the test cases passing the evaluation into test case libraries of corresponding projects. The invention has the characteristics of ensuring the quality and stability of the test case, obviously reducing the cost of manual test design and maintenance and greatly improving the overall efficiency and delivery capacity of test work.

Inventors

Guan Lingyan
SUN LINCHUN
GAO YANG
JIANG CHUN
WANG JINGXU
OUYANG XIAOGANG
WANG CAN
CHEN JIAWEI
SONG MINGLI
SUN LINJUN
SUN YUEGANG

Assignees

浙江实在智能科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260403

Claims (10)

1. Test case generation and intelligent management system based on multi-agent cooperation, which is characterized by comprising: The input and configuration module is used for interacting with a user and receiving a requirement document uploaded by the user and configuration information related to test case management; The intelligent agent cooperation module is used for generating and evaluating test cases passing the quality requirements through the cooperation of a plurality of intelligent agents based on the requirement document and the configuration information; and the RPA robot execution module is used for triggering automatic warehousing operation of the test cases after the test cases pass the review, and importing the test cases passing the review into a test case library of the corresponding project.
2. The system for generating and intelligently managing test cases based on multi-agent collaboration according to claim 1, wherein the input and configuration module specifically comprises the following processes: receiving a demand document uploaded by a user and a natural language query request input by the user; Providing an interactive form interface for collecting configuration parameters related to test case management; the interactive form includes at least the following fields: The target test management system comprises a selection field, user account information, user password information and an item identifier of the test case.
3. The multi-agent collaboration-based test case generation and intelligent management system of claim 2, wherein the agent collaboration module comprises: Generating a case intelligent agent, constructing based on a large language model, and performing semantic understanding and test gist disassembling on the required document, and automatically generating an initial test case according to a preset structured test case template; the evaluation case intelligent agent is constructed based on a large language model, and is used for automatically evaluating the initial test case according to preset evaluation rules and outputting evaluation conclusion; And the evaluation result judging module is used for triggering the generated case intelligent agent to regenerate the test case according to the trend of the control flow of the evaluation theory, if the evaluation is not passed, triggering the RPA robot executing module to execute automatic warehousing operation.
4. The multi-agent collaboration-based test case generation and intelligent management system of claim 3, wherein the agent collaboration module further comprises: the historical defect knowledge base is used for storing structured historical defect data, and the historical defect data at least comprises defect triggering scene information; And the evaluation case intelligent agent calls the historical defect knowledge base in the evaluation process, compares the initial test case with the historical defect data, and supplements or optimizes the test case according to the comparison result.
5. The system for generating and intelligently managing test cases based on multi-agent collaboration according to claim 4, wherein the process of supplementing or optimizing the test cases by the review case agents according to the comparison result is specifically as follows: When detecting that the initial test case does not cover a certain defect triggering scene in the historical defect knowledge base, automatically generating a supplementary test case for covering the scene; When detecting that the initial test case covers a certain defect triggering scene in the historical defect knowledge base, automatically improving the priority identification of the corresponding test case.
6. The system for generating and intelligently managing test cases based on multi-agent cooperation according to claim 5, wherein the generating of the test case agents specifically comprises the following processes: receiving a demand document, and analyzing function descriptions, business rules, input and output conditions and abnormal scenes in the demand document to form semantic representations of functions to be tested; Through preset prompt words, the large language model simulates the roles of the senior test engineers; generating a result by adopting a structured test case template constraint with an example; the test case template at least comprises: Use case header, preconditions, test steps, expected results, and priority fields; and generating test case data meeting the importing requirements of the test management system according to the test case template, and outputting the test case data in an Excel format.
7. The system for generating and intelligently managing test cases based on multi-agent collaboration according to claim 6, wherein the process for evaluating the test case agents specifically comprises the following steps: receiving a test case output by the generation case agent; The method comprises the steps of guiding a large language model to simulate composite roles of a senior test engineer and a developer at the same time through a preset review prompt word, and executing multi-view use case review from the aspects of testing sufficiency, realizing rationality and risk coverage; and carrying out systematic inspection on the quality of the test case according to a predefined evaluation rule system, wherein the evaluation rule system at least comprises: whether the test case format meets the requirement of the established specification or not; whether the content of the test case is complete, whether the expression is clear and has the executable performance; whether the test case covers the test scenes of functional test, performance test, safety test and boundary test.
8. The system for generating and intelligently managing test cases based on multi-agent cooperation according to claim 7, wherein the evaluation result judging module triggers the generation of the test cases by the generation agent when the evaluation is not passed, and specifically comprises the following steps: And the evaluation result judging module feeds back the evaluation opinion or the optimized prompt word output by the evaluation case agent to the generation case agent, and is used for guiding the generation case agent to conduct targeted regeneration so as to form a closed loop optimization flow of generation, evaluation, feedback and regeneration.
9. The system for generating and intelligently managing test cases based on multi-agent cooperation according to claim 8, wherein the RPA robot execution module specifically comprises the following processes: calling an automatic RPA flow of the robot flow; Filling test case data into an Excel file; Controlling an RPA executor to sequentially finish opening a webpage of a target test management system, inputting a user account password, selecting a target item and calling a test case importing function provided by the system; The test cases are automatically imported into the test case library of the corresponding item, so that centralized management and maintenance of the test cases are realized.
10. The multi-agent collaboration-based test case generation and intelligent management system of claim 9, further comprising: the abnormality monitoring module is used for uniformly monitoring abnormal conditions in the running process of the system and specifically comprises the following steps: capturing the abnormality in the execution process of each module; And executing automatic retry or abnormal termination operation according to the abnormal type.

Description

Test case generation and intelligent management system based on multi-agent cooperation Technical Field The invention belongs to the technical field of artificial intelligence, and particularly relates to a test case generation and intelligent management system based on multi-agent cooperation. Background With the continuous improvement of the scale, complexity and delivery frequency of a software system, software testing has become a key link for guaranteeing the quality and the stability of the software system. The test case is used as a core carrier for software testing activities, and the design quality of the test case directly influences the test coverage effect and the defect discovery capability. However, the following obvious technical drawbacks still exist in the process of generating and managing test cases in the prior art: (1) The design of test cases is highly dependent on manual experience, and the quality stability is poor. In the existing test flow, the test cases are manually analyzed and designed by testers based on the requirement documents, and different testers have differences in experience level, understanding capability and design habit, so that the generated test cases are difficult to keep consistent in integrity, standardization and coverage depth. (2) The automated generation capability is limited in the face of unstructured demand documents. The requirement document is usually described in a natural language form, has loose structure and various expressions, and is difficult to accurately understand the requirement semantics by using a traditional rule-driven or templated automation tool, so that high-quality test cases cannot be effectively generated. (3) Existing artificial intelligence-based test case generation schemes lack an effective quality review and feedback mechanism. Some prior art attempts to introduce artificial intelligent models to generate test cases, but usually only one-time generation is performed, automatic review, quality judgment and feedback regeneration mechanisms of generated results are lacking, and coverage of key scenes, boundary conditions and high-priority cases is difficult to ensure. (4) Historical defects are difficult to systematically multiplex with test experience. In the actual test process, a large number of discovered defects and corresponding trigger scenes, root causes and verification points are recorded in a scattered manner, and the prior art is difficult to automatically relate the history defect experience to a new test case design process, so that the known defects are easily caused to repeatedly appear in a subsequent version. (5) The application capability of the robotic process automation technology in the test case management scenario is limited. The existing RPA technology is mainly used for executing a predefined fixed flow, and can finish operations such as logging in a system, data entry, file uploading and the like. In a test case management scenario, the RPA is generally only used as a data handling and interface operation tool for importing a generated test case into a project management or test management system, and the execution process of the RPA lacks a collaboration mechanism with test case generation and review results, so that it is difficult to effectively control the imported process according to the quality state, coverage condition or review conclusion of the test case. In summary, the prior art has not provided a technical scheme capable of introducing a multi-role artificial intelligence review mechanism in a test case generation stage, dynamically supplementing or optimizing a test case by combining historical defect knowledge, and cooperatively triggering automatic execution of a robot flow after the test case passes quality review so as to realize test case generation, review and management end-to-end automation. Therefore, it is very important to design a multi-agent collaborative-based test case generation and intelligent management system which can ensure the quality and stability of test cases, obviously reduce the cost of manual test design and maintenance and greatly improve the overall efficiency and delivery capacity of test work. Disclosure of Invention The invention provides a multi-agent collaborative-based test case generation and intelligent management system which can ensure the quality and stability of test cases, obviously reduce the cost of manual test design and maintenance and greatly improve the overall efficiency and delivery capacity of test work, and aims to solve the problems of the prior art that the design process of the test cases depends on manual experience, has low efficiency and unstable quality, is difficult to continuously optimize and the like. In order to achieve the aim of the invention, the invention adopts the following technical scheme: Test case generation and intelligent management system based on multi-agent cooperation includes: The input and configuration modu