Partition based model clone detection

Model-based development has become a common development methodology in the past decade years. According to previous researches, model clone are considered harmful for software maintenance in model-based development. Existing approaches of model clone detection have a common drawback: a considerable number of detected clones are not interesting to the developers even though they fulfill model clone definition. In this paper, we present a partition based model clone detection approach, and design an experiment to evaluate the approach. The experiment result illustrates that our approach has better performance in finding more frequent and smaller clone instances which are easier to be reuse in modeling.