Abstract: The automatic software test data generation is one of the elementary problems in software testing. The problem of test data generation is reduced to a search or combinatorial optimization problem with discontinuity and nonlinearity. Therefore, it can be suitably resolved by using AI techniques such as neural networks and genetic algorithms. This paper summarizes and analyzes the application research of AI techniques on automatic test data generation. Future research directions and some of the researching problems are also addressed.

Key words: Software test, Neural networks, Genetic algorithms, Automatic test data generation

摘要： 软件测试数据自动生成是软件测试中的重要难题之一。测试数据自动生成问题可归结为测试数据的搜索或组合优化问题，通常具有不连续、不可微和非线性等特征，适合于采用遗传算法、神经网络等人工智能技术进行解决。国内外学者在此方面作了不少研究并取得一定的成果，但也存在一些问题。该文系统地综述了近年来软件测试数据智能化生成的研究和存在的问题，并对未来的发展进行了展望。

关键词: 软件测试, 神经网络, 遗传算法, 测试数据自动生成