UML模型驱动的划分测试用例生成方法研究

摘要

Generating test cases automatically is one of the main directions of the software testing research. There are a few problems such as inefficient data, no purpose, redundant, in the existing UML model-driven test methods. This paper proposed a new method to generate test case based on UML model-driven, which used partition theory to classify test scenarios into five types, and designed different ranges and ways to choose test data. And also designed a tool which realized this method. And demonstrated a tool to realize this method, which could generate test case automatically by supplied UML diagrams and data constraints of the software under test. Experimental result shows that the test cases get by this method are smaller number and targeted.%测试用例的自动生成是软件测试研究的主要方向之一.针对现有的UML模型驱动测试方法在测试数据生成方面存在低效、无目的性、冗余等问题,提出了基于UML活动图模型驱动的划分测试用例自动生成方法.该方法将测试场景归为五种类型,并为每种类型规划了测试数据取值范围和选择方法.构建了基于该方法的原型工具软件,提供被测软件原模型和数据约束即可生成测试用例.实验结果表明,该方法能够在不降低测试覆盖率的情况下,能生成数量少、针对性强的测试用例集.