嵌入式软件路径覆盖测试用例生成仿真

Simulation of Embedded Software Path Coverage Test Case Generation

为了避免因软件缺陷造成不必要的损失,提出嵌入式软件路径覆盖测试用例生成仿真。分析污点数据传播规律,定义传播模式,完成污点数据检测,探究其流向,筛选出不可行的覆盖路径;结合相似度计算公式将相似度较高的路径划分在同一组中;设计目标函数,缩小路径搜索范围,将路径覆盖问题转换为多个具有少量目标的子优化问题,建立路径覆盖数据模型;利用遗传算法,制定编码方式,确定合理的适应值,明确算法终止条件,输出最终测试用例。仿真结果证明,所提方法用例生成数量稳定,充分性强,不用占用太多的系统内存。

In order to avoid unnecessary losses caused by software defects,this paper put forward a simulation of generating test cases of embedded software path coverage.Firstly,we analyzed the propagation law of tainted data was,and defined its propagation mode.After detecting the tainted data,we began to explore its flow direction and screen out infeasible coverage paths.According to the formula of similarity,the paths with high similarity were classified in the same group.Then,we designed an objective function to narrow the range of path search,and transformed the path coverage problem into multiple sub-optimization problems with a small number of objectives,thus building a model of path coverage data.Moreover,we used the genetic algorithm to formulate an encoding scheme and determine reasonable fitness values.After clarifying the termination conditions of the algorithm,we finally output the test case.Simulation results prove that the number of case generation remain stable.And this method has strong sufficiency and does not occupy too much system memory.