Research On Path-oriented Test Data Generation

As the computer technology improves rapidly, software testing plays a nimportant role in assuring software quality and reliability. Due to the increasing scaleof software, the proportion of software testing in the whole software developmentcycle is increasing. Automated testing can effectively improve the efficiency ofsoftware testing and reduce the cost of software development, and the ke y ofautomated testing is software test data ge neration. Path-oriented test data generationis a research hotspot of software test data generation, and particle swarm opti mizationalgorithm has a good application prospect in path-oriented test data ge neration.Currently there are still ma ny shortages in the research on particle swar mopti mization algorithm and its related improveme nt, which restricts its application inpath test data generation.This paper first introduces the basic concepts of software testing and theclassification of software testing technology, and then introduces the method of testdata generation. Several intellige nt search algorithms commonly used in the test datageneration are introduced in detail, and their advantages and disadvantages areanalyzed.In order to improve the efficiency of the attractive and repulsive particle swar mopti mization algorithm, it has been deeply researched. Through a nalysis on theadvanta ges and disadvanta ges of the attractive and repulsive particle swar mopti mization algorithm, an improved attractive and repulsive particle swar mopti mization algorithm (RARPSO) which uses new formulas to calculate populationdiversity inde x and individual position vector is proposed, and several perfor mancetesting functions are used to verify its efficiency.Applying RARPSO algorithm to single path test data generation, an approach ofsingle path test data generation based on RARPSO algorithm is proposed, and it isverified by several benchmarks that the approach has higher efficiency in single pathtest data generation.Researching the issue on multi-path test data gene nration and the relatedapplications of niche technology, a new niche particle swarm opti mization algorithm(the multi-niche particle swar m optimization algorithm, MNPSO) is proposed andapplied to the multi-path test data ge neration. Then the approach of multi-path testdata generation based on MNPSO algorithm is proposed, and it is verified by several benchmarks that the approach has higher efficiency in multi-path test data generation.