Research Of Multiple-path Test Data Generation Method Based On Genetic Algorithm With Gene Repository

Test data generation is the most important part of automated software testing.For structured testing of software,data generation methods based on path coverage are widely used in software testing.In this case,the problem of generating test data evolves into a search optimization problem for test data.The most common methods used to solve this problem are the hill climbing method,the random method,the iterative relaxation method,etc.,but the execution efficiency is low and the final effect is poor.Introducing the selfinitiating algorithm into the process of software test data generation greatly improves the efficiency of test data generation.Genetic algorithms have global searchability.In the process of generating test data by using genetic algorithm,the fitness function is used as the criterion for survival of the fittest.In the cross operation of genetic algorithm,the simulated annealing process is introduced to reduce the selection of the optimal solution every time,so as to avoid convergence in the genetic process.Too fast causes the algorithm to fall into local optimum and the coverage of the path is incomplete.In the construction of the fitness function,the specific value is not added to express the path,but by calculation,the value of the value is used as a sentence covered by the full path,so that the data can be calculated by each path and then in a relatively stable interval,In the iterative process,the optimal fitness value data is added to the termination condition.In the process of finding the optimal solution by the genetic algorithm,the data corresponding to the interval of each path is recorded,so as to achieve the coverage of the full path.In the coverage process of the full path,the genetic algorithm added to the annealing process will find the optimal solution faster,skipping the path with a slightly lower fitness range,resulting in incomplete coverage of the full path.To avoid this problem,In the whole genetic process,a static gene pool is introduced to ensure the diversity of the population.After the genetic algorithm completes a generation of cross operations,the individuals are sorted,and the static gene pool is introduced to be more dispersed than the current generation.Thereby increasing the diversity within the population and increasing the generation of different fitness interval data in the iterative process.Thus,based on the genetic algorithm,an annealing process is introduced,and a static gene pool is added to achieve full path test data generation.Through establishing an annealing genetic algorithm model that joins the static gene pool,the full-path output test is performed on the more common typical algorithm codes,and the algorithm model is not sensitive to the adjustment of each parameter.Compared with the full-path test data generation method used in recent years,the hybrid algorithm model has greatly improved the effect of test data generation,and the number of iterations is also faster.