Automatic Test Data Generation Based on SA-QGA
Haijing Ji, Junmei Sun · 2012
Recently, genetic algorithm and their evolutionary algorithms are widely used on automatic test data generation, but they have many problems such as local optimum, premature convergence and being difficult to find global optimum. This paper proposes a new algorithm: SA-QGA (simulated annealing - quantum generate algorithm), and introduces the Boltzmann mechanism of SA into B_QGA (QGA based on Boltzmann selection mechanism). The SA-QGA is used to generate test data with the excellent computing and global search performance of QGA and local search capability of SA. The experiment result verifies that SA-QGA has good performance on test data generation.