LAUSR

Abstract Searching based testing case generation technology converts the problem of testing case generation to function optimizations, through a fitness function, which is usually optimized using heuristic search algorithms. The particle swarm optimization (PSO) optimized testing case generation algorithm tends to lose population diversity of locally optimal solutions with low accuracy of local search. To overcome the above defects, a self-adaptive PSO based software testing case optimization algorithm is proposed. It adjusts the inertia weight dynamically according to the current iteration and average relative speed, to improve the performance of standard PSO. An improved alternating variable method is put forward to accelerate local search speed, which can coordinate both global and local search ability thereby improving the overall generation efficiency of testing cases. The experimental results demonstrate that the approach outlined here keeps higher testing case generation efficiency, and it shows certain advantages in coverage, evolution generation amount and running time when compared to standard PSO and GA-PSO.