基于占优度与集合进化的并行程序变异测试数据自动生成

Automatic generation of test data is the premise of software testing automation which can improve the efficiency of testing. Mutation testing has strong capabilities in detecting faults, many severe problems existing in previous mutation testing methods, however, cause these methods difficult to be applied in the testing of concurrent programs. Aiming at the characteristics of message-passing concurrent programs, this project studies the issue of automatic generation of test data for mutation testing of concurrent programs based on the dominant degree and set evolution, and expects to obtain the following achievements: (1) presenting the theory of mutant reduction based on the dominant degree, greatly reducing mutants obtained from performing mutation operators with different kinds on different positions; (2) constructing the mathematical model for the problem of generating test data for mutation testing based on path coverage, effectively reducing its difficulty; (3) designing the method of automatically generating test data for mutation testing based on set evolution, further improving the efficiency of test data generation; (4) applying the proposed theory and method in the mutation testing of concurrent programs, significantly enhancing their reliability. The project, at the cross edge of such disciplines as automation, computer, and mathematics, is novel and challenging, and is urgently demanded by industry. The achievements of this project can enrich the theory of mutation testing, improve the efficiency of testing concurrent programs, and extend the scope of applying set evolution optimization. Therefore, the project is of considerable significance in theory and application.

软件测试自动化能够提高测试效率，自动生成测试数据是软件测试自动化的前提。变异测试具有很强的检错能力，但是，已有方法存在的许多严重问题，使其难以应用到并行程序测试中。本项目针对消息传递并行程序的特性，研究基于占优度与集合进化的并行程序变异测试数据自动生成。通过研究，拟提出基于占优度的变异体约简理论，大幅度约简在不同变异点，通过实施不同类型变异算子得到的变异体；建立基于路径覆盖的变异测试数据生成问题的数学模型，有效降低问题求解的难度；设计基于集合进化的变异测试数据自动生成方法，进一步提高测试数据生成的效率；将所提理论与方法应用于并行程序变异测试中，显著提高程序的可靠性。本项目是自动化、计算机，以及数学等学科有机交叉、新颖且富有挑战性的研究方向，有非常明确的产业需求；产生的成果能够丰富变异测试理论，提高并行程序测试效率，并扩大集合进化优化的应用范围，因此，具有重要的理论意义和实际应用价值。

软件测试自动化能够提高测试效率，自动生成测试数据是软件测试自动化的前提。变异测试具有很强的检错能力，但是，已有方法存在的许多严重问题，使其难以应用到并行程序测试中。项目负责人及成员对基于占优度与集合进化的并行程序变异测试数据自动生成，进行了为期4年的全面深入研究，提出了基于占优度的并行程序变异体约简理论，建立了基于路径覆盖的变异测试数据生成问题的数学模型，设计了基于集合进化的变异测试数据自动生成方法，并将所提理论与方法成功应用于并行程序变异测试中。基于上述成果，获高等学校科学研究优秀成果奖自然科学二等奖1项、授权发明专利5项、计算机软件著作权登记4件，出版专著2部，在被SCI或EI等检索的国际国内学术期刊上发表论文25篇，其中，SCI源刊论文13篇（含顶级期刊论文2篇，中科院1、2区期刊论文4篇），培养博士研究生3名、硕士研究生6名，圆满完成了项目的预期目标。研究成果解决了并行程序变异测试数据难以自动生成的难题，丰富了变异测试理论，提高了并行程序测试效率，扩大了集合进化优化的应用范围，具有重要的理论意义和实用价值。