Research On Combinatorial Test Case Generation And Optimization For Fluid-Structure Coupling Simulation Software

During the use of fluid-structure coupling simulation software,system faults caused by the interaction of parameters often occur.As the fluid-structure coupling simulation software system tends to be large and complicated,the space of software parameter combination increases explosively,and it is difficult for the test case set to cover all combinations exhaustively.The rapid generation of combinatorial test cases with high coverage and small scale is an important technical challenge for fluid-structure coupling simulation software testing.However,the current combinatorial test case sets based on algorithms or manually generated have a large number of redundancy problems in the case of full coverage combination.Therefore,it is of great practical significance and application value to study automatic test case generation and optimization method of fluid-structure coupling simulation software and realize automatic test platform.Aiming at how to simplify the size of test case set as much as possible while meeting the requirement of combination full coverage,this paper studies from two aspects.On the one hand,this paper proposes an improvement to the existing test case generation method to generate a more streamlined set of combinatorial test cases.On the other hand,this paper further optimizes the generated test case set and effectively reduces the number of combinatorial test case sets.The main research contents of this paper are as follows:(1)In order to solve the problem that genetic algorithm tends to fall into local optimality in the process of generating composite test cases,which leads to excessive redundancy of the generated combinatorial test cases,the genetic algorithm is improved.Firstly,in the iterative evolution process of genetic algorithm,the idea of mountain climbing method is introduced,and the optimal individual of each generation is directly put into the next generation,so as to avoid the destruction of cross mutation steps and improve the stability of genetic algorithm.Secondly,a population split migration strategy is proposed to dynamically adjust the evolutionary direction and search range of the population,and improve the generation ability of genetic algorithm’s compact combinatorial test case set.Finally,environmental mutation steps are added after population migration to avoid local optimization and improve the probability of generating high-quality test case sets.The results show that the improved algorithm can obtain a smaller test case set compared with PICT,Allpairs,ACTS and other combinatorial test tools,genetic algorithms and variant algorithms.(2)Aiming at the problems of large scale and high redundancy of the existing combinatorial test case set,an optimization method is proposed.Firstly,a computational method of parameter combination redundancy is proposed to deeply mine the redundancy information of the test case set.Secondly,the test cases are sorted from high to low by the number of real combinations,the real combinations of the lower test cases are extracted,inserted into the upper test cases,and the virtual test cases are deleted,so as to realize the simplification of the test case set.The experimental results show that the optimization algorithm in this paper can effectively reduce the size of test case set generated by genetic algorithm and provide method support for testers to optimize the existing test case set.(3)Aiming at the testing requirements of fluid-structure coupling simulation software NNW-FSI,a combinatorial test case automatic generation and optimization tool is designed based on the proposed improved genetic algorithm and parameter combinatorial coverage redundancy optimization method.Firstly,the characteristics of fluid-structure coupling simulation software are analyzed.The requirements analysis of the tool determines that the tool should include the functions of parameter data generation,test case generation,optimization,management and evaluation,and the overall design and detailed design.Secondly,the test tool is implemented by programming using the APP Designer module of the MATLAB language.Finally,the functional verification is carried out on Rae examples to prove the effectiveness and utility of the tool.