基于遗传算法优化的SOFM神经网络生成测试数据集的方法

Method of Generating Test Datasets Using Genetic Algorithm Optimized SOFM Neural Network

智能算法正成为软件测试领域新兴研究方向,运用智能算法生成复杂软件的测试数据已成为一种广受推崇的方法。采用基于遗传算法的技术生成测试数据,能够生成满足测试覆盖要求的少量测试数据。然而,对于生成大量测试数据集的情况来说,该方法并不适用。为了能够快速生成满足测试覆盖要求的数据集,提出一种基于遗传算法优化的自组织特征映射(SOFM)神经网络生成测试数据集的方法:首先,利用遗传算法的全局搜索能力,从海量数据中筛选出少量满足测试覆盖要求的代表性数据。接着,以这些遗传算法生成的测试数据为基础,结合SOFM神经网络强大的侧向联想能力,旨在生成大量满足测试覆盖要求的测试数据集。实验结果表明,该方法有效提高了测试数据集生成的效率。

Intelligent algorithm is becoming a new research direction in the field of software testing.It has become a widely respected method to generate complex software test data using intelligent algorithm.Using genetic algorithm based technology to generate test data,it can generate a small amount of test data to meet the requirements of test coverage.However,this approach is not suitable for situations where large test datasets are generated.In order to quickly generate data sets that meet test coverage requirements,a method of generating test data sets based on self-organizing feature mapping(SOFM)neural network optimized by genetic algorithm is proposed.Firstly,a small amount of representative data meeting test coverage requirements is selected from massive data by using the global search capability of genetic algorithm.Then,based on the test data generated by these genetic algorithms,combined with the strong lateral association ability of SOFM neural network,the aim is to generate a large number of test data sets that meet the requirements of test coverage.Experimental results show that this method can effectively improve the efficiency of generating test data sets.