Model transformation using logical prediction from sequence diagram: an experimental approach

Recent trends on software development life cycle (SDLC) deal much on automatic processes that leads to time and cost reduction. In the era of model driven architecture (MDA), unified modeling language (UML) models are the backbone of any developing software. A minimum of 5 models out of 14 models need to be designed to completely visualize any software. The proposed logical prediction model transformation automates the transformation of two models from the sequence diagram. Each model carries same information of the other models in different aspects to visualize the requirement constraints in different dimensions. In order to take the advantage of this, sequence diagram has been considered as a pioneer from MDA approach to generate other models automatically. Information such as elements, attributes, relationships, etc., of the sequence diagram are extracted using XML object model parser. The extracted information from sequence diagram combined with the pre-defined logical prediction rules, generates the elements and relationships of other models. The outcome of transformed information is rendered by PlantUML structure to produce the desired model. The experiment undertaken has been focused to generate class diagram and activity diagram based on the pre-defined logical prediction rules. It also confirms the transformation process proposed, generates suitable and appropriate class and activity diagrams. This transformation process has a scalability to generate any model from the other model based on proper logical prediction rules. This automation proposal eases the task of designer in design engineering phase of the SDLC.