Automatic generation of software test data based on hybrid particle swarm genetic algorithm

A hybrid particle swarm genetic algorithm is purposed to apply in software testing using case automated generations. On the basis of classical genetic algorithm, the algorithm divided the population into “families”, influencing the convergence efficiency by crossover in family, keeping the diversity of the population by crossover between families; meanwhile, enhancing the speed of convergence by the PSO crossover (commixed the thought of PSO in genetic algorithm) According to the characteristics of software testing problems, we designed the corresponding fitness function and the encoding method. The results of data experiment were given to illustrate the effectiveness of the algorithm.