Study On Automatic Generating Path Test Data Based On PSO

Generating path test data is one of hot points in the research field of softwrare test. Automatic path test data generation is vital to improve the efficiency of path test and reduce test cost. Particle swarm optimization (PSO) is employed to generate automatically path test data in this thesis. The contents of this thesis are shown as follows:⑴Single path test is a basic software test technology, a method based on PSO is proposed for automatic generation of single path test data. Proposed method uses the sum of branch functions to construct the fitness function. The experimental results, which compare proposed method with the genetic algotirhm based method, indicate that proposed method is superior to the compared method, and its performance is more stable.⑵Automatic all paths test data generation of a program is challenged. Based on PSO, a method of generating all paths test data of a program is presented. A new fitness function is constructed, and the frequencies for all paths are registered. Generation results show that the efficiency has the enhancement of all paths test data generation compared wtih single path test data generation.⑶Based on above research results, a tool of automatic generating path test data is developed by Java. It is described detailedly in this thesis.The test results show that the tool is practical and effective, can generate automatically the path test date for C programs.