Research on Intelligent Software Testing Architecture and Key Technologies Based on Multi Agent Collaboration

Abstract

With the exponential growth of software system complexity and continuous compression of delivery cycles, traditional software
testing methods that rely on manual experience and static scripts are no longer able to meet the requirements of rapid iteration in terms of
efficiency, coverage, and response speed. This study aims to systematically explore how artificial intelligence technologies, especially big
language models, machine learning, and deep learning, can deeply empower the entire software engineering testing process and achieve a fundamental paradigm shift from "automation" to "intelligence". The study first conducted an in-depth analysis of the current development status
of intelligent testing and its core challenges in data quality, model generalization, and engineering implementation. On this basis, a hierarchical decoupling and modular intelligent testing optimization technology framework was constructed, and its core methods in key aspects such
as intelligent generation of test cases, defect risk prediction, execution scheduling, and result analysis were elaborated in detail.