Studies On Generating Of UML Behavior Models In Reverse Engineering

For the complexity of software systems, the main works of software development, maintenance and evolution have been focused on the comprehension of existing systems. Reverse engineering is an effective technique for program comprehension. The aim of reverse engineering is to identify the components and their relationships, create representations of the system in another form or at a higher level of abstraction.On the basis of profound analyses of reverse engineering, UML2.0 and XML2.0, this thesis mainly focuses on the methods of generating UML behavior models in reverse engineering. First of all, an approach to create the UML sequence diagrams reversely using dynamic information combined with static information is brought forward. The sequence diagrams gained by this method are comprehensive and compatible with the UML2.0 standard. Then the state machine is also future researched with the help of UML2.0 sequence diagrams. An algorithm, which is improved from the BK algorithm, to generate a statechart diagram is proposed. The statechart diagram generated by this algorithm is closer to the actual model, which is more useful for the user to understand the target system generally.All of the algorithms and methods provided in this paper have been implemented in a toolkit named XDRE( XiDian Reverse Engineering).