Research And Application Of Traceability Link Recovery For Software Requirement Relationship Networks

With the rapid development of the Internet industry in recent years,the number of related software products has also increased exponentially.Software products have short development cycles but high quality requirements.In the face of complex market environment and changing user requirements,requirements often need to face the risk of change.In requirement engineering,the role of traceability links is to achieve safe and reliable change operations to ensure the consistency between requirements and products.It is becoming increasingly important to better establish traceability relationships between requirements and codes to improve the speed and quality of software development.Requirement change management system is studied in this thesis.The system contains three core modules: requirement management module,requirement change analysis module and user management module.This system uses the EAV(Entity-Attribute-Value)database model to provide basic data storage services,on the basis of which requirement traceability links are automatically established and basic management operations of requirement changes can be realized,aiming to ensure the consistency of requirements and products when changes are made,thus improving the development efficiency of software products.This thesis focuses on the task of traceability link recovery for software requirement relationship networks with the following main research components:This thesis uses the EAV database model to provide data support for the system.Requirement documents organized from requirements data are non-relational and these requirements are often accompanied by different attribute information.However,nonrelational databases do not have sufficient support for transactions.In this thesis,the entity-attribute-value of requirements are stored separately,and the separate data row operations are decomposed into multi-step operations on the entities,attributes and values of requirements,while the speed of requirements loading is improved by caching.Using this database model solves the problem of storing requirements in relational databases facing large information redundancy and the security risk of dynamic operation of database.A kind of requirement traceability link recovery algorithm is designed for better automatic establishment of traceability links between software requirements and source codes.By combining fine-grained analysis and query expansion techniques,the structural information inherent within the requirements and source codes is extracted into finegrained requirements and code elements.And considering the more obvious correspondence between software artifacts caused by the fine-grained granularity,the problem of establishing traceability links for fine-grained artifacts is modeled as a retrieval problem.The requirements to be matched are corrected using query expansion technique and then queried twice.The results of the query are aggregated through an aggregation step and a double-threshold strategy is used to obtain the final traceability links.Compared with traditional methods,the algorithm proposed in this thesis further improves the effectiveness of software requirement traceability link recovery.In addition,requirement change management system is designed in this thesis in order to manage requirements when software requirements are changed.Requirement files that require requirement traceability link recovery can be imported into the requirement change system and requirement entities and attributes can be edited.At the same time,traceability links can be established automatically by the system.The requirements with traceability links can be set as the baseline after the requirement manager makes manual corrections,so that the code with the traceability relationship can be modified synchronously through the traceability links when the requirements are changed to achieve safe requirement changes.