Research And Implementation Of The Auto-mapping From Requirement Model To Business Model

In the field of software production, System's requirements analysis is the basis of software production. Therefore business model should correspond with system's requirements analysis in order to ensure the software quality. How to quickly and correctly modify business model according to the changes of requirements, has became one of the hotspots in the software research domain. In the magic platform, developers use a visualization platform to customize software following requirement documents. But this artificial customization contains some questions, for example it is probably easy to make a mistake for developer to comprehend requirement documents and business customization is complicated. All of these questions become to the customization bottleneck. Moreover, this customization way can't make dynamic changes when requirement changes.In this thesis, based on MDA and model transformation technology, it discusses a transformation method from requirement model to business model, and provides a solution for business process refinement. This research includes the following aspects:Firstly, this thesis makes formal description for magic platform requirements analysis, analyses element relationship among requirement meta-model. According to the characteristics of e-government, it improves the model description method based on XML using process-oriented model building approaches. Then it uses this method to describe requirement meta-model and requirement model.Secondly, this thesis discusses a method to describe a domain-oriented target process meta-model. In connection with business modeling process of magic platform, it conceives a business meta-model of magic platform, and the meta-model is described using UML.Finally, this thesis puts forward the architecture of auto-mapping system, and carries on design and realization for each part of the system structure. It emphatically analyzes on transformation rule library in model transformation engine, and makes formal description for the rules. It designs four kinds of scheduling algorithms in model scheduling library, and depicts model transformation process under the control of model process manager. At last, the automated mapping system is verified by an instance, and is analyzed in respect of the system performance.