Research On Second-order Mutant Reduction Based On SOM Neural Network Model

Mutation testing is software testing technology by manually injecting one or more defects into the original program to improve the quality of test cases.Higher-order mutation testing simulates the actual complex defects in the original program by manually injecting two or more defects into the original program,which is of great significance in the mutation testing.In the existing research on mutation testing,most of them are based on second-order mutation.However,the number of second-order mutants formed by the combination of first-order mutants will greatly increase,which will bring large mutants execution costs and improve the difficulty to find second-order mutants increasing test cases,ultimately increase the execution overhead of the second-order mutation testing.In order to reduce the execution overhead of the second-order mutation testing in the running procedure,this thesis proposes a method of second-order mutant reduction based on SOM neural network.Our proposed method firstly utilizes a more comprehensive combination strategy to generate feasible second-order mutants based on traditional first-order mutant generation.And then based on simple intuitive factor analysis and complex factor analysis in a variety of situations,various factors affecting the similarity of intermediate values in the execution of second-order mutation testing are determined to construct accurate SOM neural network.And at last mutants are clustering based on such model to achieve execution cost of mutation testing reduction.In order to verify the soundness and effectiveness of our method,theoretical analysis is first carried out,and then the example verification is based on the classic benchmark and the open source projects.The analysis results of mathematical theory show the soundness of our method.Experimental results show that on the one hand,the use of a more comprehensive combination strategy can fully simulate the complex defects of the programs.On the other hand,the number of mutants has decreased significantly and subtle second-order mutants that facilitate the addition of test cases have been found by our method,thereby achieving the goal of reducing the execution overhead of the second-order mutation testing.