Research On The Key Technology Of Model Transformation Based On Ontology Semantic

Model-Driven Architecture, MDA is proposed by the OMG organization in March, 2001 which is one new software architecture methodology. Its core idea is to abstract the Platform Independent Model, short for PIM, which has nothing to do with the platform and concert technology, then in view of different realization platform to draw up transformation rules, according the mapping rules, the model transformation engine transforms PIM to Platform Specification Model, shor for PSM, which realizes with the specific platform and technology. And finally, the code generator converts PSMs to the target code, which realizes the software development process. MDA solves the complexity of software development by high abstraction level, reduces the impact of changing demands on the implemention by separating the business logic with implementation technology by the greatest degree.Then with the aid of model reuse and model transformation,extending the software life cycle, enhances the efficiency and quality of software development.Model transformation is the core technology of MDA, which throughout the process in the model driven development. This paper takes model transformation as the research object, in-depth studies the method of model transformation, and proposes a model transformation framwork that is based on ontology. Through the improvement of the current model transformation in semantics uniform to obtain a more perfect model.In general, the main contributions of this paper are listed:The thesis proposes a semantic ontology-based model transformation mechanism, which is propsed to solve the lack of the semantic characterization among model transformation.By this way, the semantic characterization for model driven development can be ensured.The thesis founds the semantic ontology library, which is based on semantic ontology model transformation mechanism.The whole ontology library is divided into threelayers:abstraction layer, concrete layer and instance layer.The abstraction layer is the basic element of the basic concepts, structural elements and characteristics of elements. In order to enhance the scalability of the concept, semantic ontology database defines the feature set of elements, including property descriptions and constraints, which provides semantic database support for developers.The thesis sets rules of ontology-based semantic model transformation, which defines the mapping between source model and target model elements. According to the mechanism of ontology-based semantic model transformation, the transformation rules are divided into three parts:syntax transformation, semantic transformation, transformation of binding syntax and semantics, the rules is made corresponding to different abstraction levels, so that the model transformation rules are guided.By exploring the critical techniques in actual field, an example of model-driven software development is given to show the feasibility and validity of the proposed mechnism.