Research On Multi-objective Testing Resource Allocation For The Architecture Based Software Model

In modern software engineering,software system is usually developed by selecting suitable and ready-made reusable components,and then assembling these components with the clear software architecture.Because this technology can significantly reduce the cost and time of software development,it has been widely used in the actual software industry.However,software testing consumes almost half of the software development resources.Therefore,how to allocate testing resources reasonably,so as to consume as few testing resources and testing costs as possible and achieve as high software reliability as possible,has always been a hot and difficult topic in the field of software engineering.Particularly,in recent years,the western countries headed by the United States are intensifying the technology blockade on our country,which leads to the increasingly prominent problem of our industrial software "being hit in the throat".The industrial software,which is related to independent and controllable industries,has become one of the key areas in the new development pattern.How to improve R & D efficiency of domestic industrial software has become one of the current opportunities and challenges.For the testing resource allocation problem,most of the existing research is limited to the parallel-series modular software model,which does not fully consider the important system characteristics of the system architecture(depending on the operation profile in practical application),over relies on the performance of the search algorithm,and does not consider the high requirements of the software project manager for software reliability,which is difficult to meet with the practical application.Consequently,this dissertation studies the reliability-constrained multi-objective testing resource allocation problem for the architecture based software model.The main research work of this dissertation is summarized as below:(1)The research background of software testing resource allocation is introduced,and the related work and the existing problems are discussed.(2)The architecture based software model,the reliability and cost evaluation model,the multi-objective optimization technology,and several classical multi-objective evolutionary algorithm are introduced in detail.(3)The multi-objective testing resource allocation problem under reliability constraints for the architecture based software model is presented.Then,to achieve the required reliability,the new upper and lower bounds of testing time invested in each component are derived by using the Lagrange multiplier method and the dichotomy method,which greatly reduces the solution space.The experimental results show that the new upper and lower bounds of the decision variables derived from the knowledge of the problem itself can shift the infeasible solutions into the feasible region,so as to drive the whole population to always evolve in the feasible region,which can significantly enhance the search performance of the multi-objective evolutionary algorithm,and obtain more explanatory and higher-quality solutions.(4)A multi-stage and multi-objective testing resource allocation model is first constructed based on the architecture-based software model and the dynamic changes of reliability and fault number.Then,a multi-stage and multi-objective testing resource allocation algorithm under dynamic reliability and fault number is proposed on the basis of parameter re-estimation,population re-initialization,the generalized differential evolution,and weighted normalized sum.The experimental results demonstrate that the proposed algorithm can better adapt to the dynamic changes of reliability and fault number and can provide more and better testing resource allocation schemes for the dynamic testing of the architecture-based software model.