Test Data Generation for Multiple Paths Based on Local Evolution

作者：Yao Xiangjuan^*; Gong Dunwei; Wang Wenliang

来源：Chinese Journal of Electronics, 2015, 24(1): 46-51.

DOI：10.1049/cje.2015.01.008

• Test data generation
• Path coverage
• Genetic algorithm
• Local evolution

Generating test data by genetic algorithms is a promising research direction in software testing, among which path coverage is an important test method. The efficiency of test data generation for multi-path coverage needs to be further improved. We propose a test data generation method for multi-path coverage based on a genetic algorithm with local evolution. The mathematical model is established for all target paths, while in the algorithm the individuals are evolved locally according to different objective functions. We can improve the utilization efficiency of test data. The computation cost can be reduced by using fitness functions of different granularity in different phases of the algorithm.

• 出版日期2015-1
• 单位中国矿业大学（北京）