A Novel Optimized Path-Based Algorithm for Model Clone Detection

Abstract

According to previous reports software clones are considered harmful for software maintenance. Likewise, model clones are problematic in model-based development. It is significant to detect model clones in software models. In this paper, we present a novel optimized path-based model clone detection algorithm (OPMCD). It first builds paths from block graphs, and then identifies clone instances from the common subsequence of paths. Moreover, an experiment is designed to evaluate the algorithm through comparing with the state-of-the-art of model clone detection algorithm ConQAT model clone detection (CMCD). The experiment result illustrates that OPMCD has better performance in terms of efficiency, and it is practically suitable for large-scale MATLAB/Simulink models.

Duplicated code pattern mining in visual programming languages

“…To the best of our knowledge, ours is the first graph-based approach that solves the scalability issue by means of an inverted index. Some approaches exist in the literature for detecting duplicated code in Simulink models [1,13,29,41,50]. SIMONE [1] applies the NiCaD [11,15] text-based detector on textual representations of the models, thus sacrificing visual structure.…”

“…Such patterns allow the developer to understand and address the sources of code duplication more effectively. Alternatively, some graph-based algorithms for text-based languages [26,30,53,63] or other VPLs [1,13,29,41,50] could be directly applied to Out-Systems logic, but these typically suffer from scalability issues due to the hardness of checking sub-graph isomorphism, and the ones that do address this issue perform some approximated form of sub-graph matching, thus not guaranteeing the consistency of the graph structure.…”