软件模糊测试中遗传杂交算法的研究  

Research on genetic crossbreed algorithm in software fuzzy test

作　　者：沈亚楠[1,2] 赵荣彩[1] 王小芹[2] 任华[2] 庞建民[1] 李大伟[2] 

机构地区：[1]信息工程大学信息工程学院,郑州450001 [2]北京北方计算中心,北京100091

出　　处：《计算机应用》2009年第B12期141-142,共2页journal of Computer Applications

摘　　要：利用遗传算法的导向性原理,可以不断寻找能够触发软件异常的测试用例。杂交过程是遗传算法中重要过程之一,如何定义父代中的优秀基因并遗传到子代,对寻找算法搜索空间的最优解非常重要。定义了重要基因及优良选择的概念并提出相应杂交算法。在Windows RPC测试中发现,该算法能够寻找到种群中优良基因并不断遗传到子代中,为自动构造软件输入数据,触发软件异常提供支持。Fuzzing test based on genetic algorithm is a very foreground automatic test technology. Based on the oriented principle of the genetic algorithm, the test data which can trigger software exception will be found ceaselessly. Crossbreed is one of the important processes in genetic algorithm. How to define father＇s best gene and descend to the child is very important to find the best solution in algorithm searching space. Important gene and its best choice were defined and the corresponding crossbreed algorithm was designed. In Windows RPC test, best gene can be found and be passed down to the next generation using the proposed algorithm. Automatic software input data is constructed and triggering software exception is supported.

关 键 词：模糊测试 遗传算法 杂交算法 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]