Optimal Testing Strategy For Multi-stage Software Engineering Project Considering Compound Cost

In recent years,with the continuous improvement of hardware performance,various emerging software R & D and testing technologies have emerged.The continuous development of big data,cloud computing,distributed system architecture with high concurrency and low delay and artificial intelligence has injected strong impetus into the software industry,prompting software practitioners to develop more efficient and convenient software application systems.With the increasing volume and complexity of the software application system itself,the errors and defects hidden in the software system increase exponentially,and the requirements for the robustness and reliability of the complex software system are more strict,which brings unprecedented risks and challenges to the software testing work.In the case of limited external resources,how to efficiently and quickly test various defects in the software and correct them in time is the biggest difficulty faced by various software enterprises and practitioners.Traditional software testing resource allocation methods have been difficult to meet the testing needs,so software testing resource allocation and cost estimation have gradually become the focus of software engineering project managers.How to more truly model and describe the software test process and accurately allocate test resources according to the test process to reduce the test cost has become an important problem faced by managers,which is also the research direction of this thesis.The main research contents and innovative work of this thesis include:(1)A nonlinear functional relationship is constructed to describe the relationship between software testing resource investment and system defects found in testing.Two nonlinear function expressions are constructed through two test resource decision variables: the allocation proportion of software testers and the allocation of test time in each stage,and the relationship between the decision variable expression and the test cost before software release and the loss cost caused by software legacy defects after software release is analyzed,A software testing resource allocation model with compound cost as the objective function is established.(2)Using formula derivation and numerical solution,the decision variables of software test resource allocation model are solved and analyzed,and the module test efficiency and The impact of the changes of three exogenous variables,the estimated number of defects in the module and the cost of repairing each defect,on the two decision variables,the time allocated to each stage of the test,the proportion of testers allocated to each module in the test stage,and the total cost of the test.(3)Based on the test process of a commodity management system,a multi-stage and multi module software test resource allocation model is constructed,the exogenous variables are assigned and brought into the model,and the optimal test resource allocation scheme and the optimal solution of software test compound cost are obtained by using genetic algorithm.On this basis,set the value range of various exogenous variables in the multi module and multi-stage software resource allocation model,observe the impact on the model decision variables through the changes of exogenous variables,and analyze the causes of this impact.Through the research,analysis and practice of multi-stage and multi module software test resource allocation model,this research can provide more accurate and reasonable test resource allocation scheme for relevant decision-makers of software engineering projects,and provide reference basis for the overall cost estimation of software testing.