测试用例自动生成方法的现状及研究

软件测试数据生成在软件系统开发费用中占很大比重。如果该过程能自动实现,则会极大地减少软件开发的周期和费用。测试用例的生成工作包含选定被测任务、分析输入数据、确定其取值并分析对应的输出数据。其中分析对应的输出数据是决定测试是否成功的关键环节。测试用例选取的一个中心原则,就是以用最少的测试用例找到尽可能多的错误。目前的工具尚不能完成自动生成测试用例这个环节,往往是只能采用人工选取的方法。按所采用的方法和研究对象的不同,将测试用例自动生成方法主要分为5类:基于有限状态集的测试,基于标注的转换系统的测试,针对面向模型的需求规格说明的测试,针对面向对象软件的测试,以及运用模型检查生成测试用例的方法。在简单介绍前4种方法之后,重点对模型检查的方法进行详细的分析和探讨。

Research on Status of Automated Test Case Generation Method

It costs much money that software test data generation in software development.If this process can realize automatically,it will reduce period and money a lot in software development.Test case generation includes of choosing the object tested,analyzing the data input,making sure the value of the data and analyzing the relevant data output.It is key of analyzing the relevant data output that decides whether test is successful or not.The central principle of choosing test case is that it can find more error using less of test case.At present,there are few tools that can generate test case automated completely.Now,people always do it man-made.Based on the difference of the methods and object researched,it has five methods of automatically testing case generation.It includes that test based on finite state,test of transformation system based on label,test aiming at model-oriented requirement specifications,test aiming at object-oriented software and method of automated test case generation by model checking.After some brief introduction of the four methods before,it gives more information about the method of model checking here.