The Research On Automatic Generation Of Test Data Based On Intelligent Optimization Algorithm

Software testing can ensure the quality of software,but it also costs almost half of the software development.Traditional software testing usually requires manual design the test data,but as the software scale increasing,manual design the test data can't be satisfactory.Therefore,Automatic generation of test data has become a research direction,it can reduce the costs and time of soft development and improve the efficiency of soft testing.The automatic generation of test data includes not only automatic generation of test data for single path but also for multipath.To improve the efficiency and coverage of test data generation,this paper has done the following two parts of the research:(1)In this paper,a new algorithm for automatic generation of test data for single path is proposed,the dynamic increasing compression factor is added in the Firefly Algorithm,and proposed the DCFA(Dynamic Compression Firefly Algorithm).Firefly algorithm is a novel intelligent optimization algorithm,and it is simple to realize,fast convergence speed,it has great development potential in single path test data automatic generation.But the FA has some deficiencies,such as the weak balance between its local search ability and its global search ability,in order to solve this problem,dynamic inertia weight is added in this paper.In the local search,FA uses the strategy of random search,in order to enhance the local search capability of the algorithm and control the search area of FA,in this paper,the dynamic increasing compression factor is added in the FA,and proposed the DCFA,then it is applied in the automatic generation experiment of the test data oriented to the single path.The experiment proves that it can find the optimal solution faster and generate the corresponding test data.(2)In the research of automatic generation of test data for multi-path,FWPSO(Fireworks Particle Swarm Optimization)algorithm is proposed in this paper.it is based on Particle Swarm Optimization and Fireworks Algorithm.The difficulty of multipath testing is how to generate data that can cover multiple targets in a single search,at the same time,the calculation of similarity between paths are also key points.In this paper,the importance of different factors affecting path similarity is considered in the design of path similarity calculation method.This paper chooses to apply Particle Swarm Optimization to automatic generation of multipath test data,but considering the late running of PSO,particle diversity is reduced andis easy to fall into local extreme.In this paper is added the idea of explosion operator in the algorithm of Fireworks Algorithm,and proposed the FWPSO based on Fireworks Algorithm.At the end of each iteration,the current optimal particle "explodes" to produce a new particle,and then compares the selection of the optimal particle for the next update iteration.This can not only increase the diversity of particles,but also help the algorithm jump out of the local extreme and find the global optimal solution.Finally the proposed FWPSO algorithm is applied to the multipath test data generation experiment.Experimental results show that the algorithm improves both in optimization efficiency and coverage rate in multipath testing.