Research On Traceability Recovery And Maintenance Between Requirements And Code

Software traceability refers to the ability to create and maintain relationships(also known as trace links)between artifacts(such as requirements and code)during the software development process,and use these relationships to perform a series of analyses on software projects.Software traceability supports activities such as change impact analysis,dependency impact analysis,system verification,and safety certification.In general,most research related to traceability is conducted on requirements and code.At present,there are two main problems in traceability recovery and maintenance between requirements and code.On the one hand,when the number of requirements and code elements is large,the creation of trace links requires a lot of labor costs(sometimes even more than the programming implementation itself),and the quality of traceability is difficult to guarantee;on the other hand,when a software system that has achieved traceability evolves,the requirements and code in the system may change.As a result,developers and maintainers need to maintain the achieved traceability dynamically based on these changes.However,due to frequent updates of the software system,it is difficult to update the affected trace links in a timely manner.In view of the above problems,this thesis proposes a method for automated traceability recovery and maintenance between requirements and code.This method can automatically create trace links between requirements and code,and can automatically identify outdated requirements during the process of system evolution.The research contents of the thesis include:(1)Automated traceability recovery: A method for creating trace links between requirements and code using active learning is proposed.First,a set of training features that can reflect the characteristics of trace links is defined;Then,the steps of training a classifier based on an active learning algorithm are explained;Finally,the classifier is used to identify trace links between requirements and code.(2)Automated traceability maintenance: A method for identifying outdated requirements based on code change analysis is proposed.First,a set of rules to detect changed code elements that affect requirements is defined;Then,the way to track changed code elements to requirements is introduced;Finally,a list of outdated requirements is generated.(3)According to the method proposed in this thesis,a system is designed and developed that can create trace links and identify outdated requirements.Finally,a case study is used to demonstrate the running process of the system,illustrating the feasibility and effectiveness of the method in this thesis.