Similarity Analysis of Control Software Using Graph Mining

The control software of large scale industrial systems such as machines and plants in the domain of automated Production Systems (aPS) is oftentimes developed using the method clone and own. However, cloning is one of the reason for high maintenance cost in software lifecycle and thus, the need for clone detection is arising. In large scale control software, clone detection is a tedious work, which cannot be performed manually. In order to alleviate the problem of huge number of clones in control software, structural clone detection can be performed. Detecting structural clones can help in better understanding of large scale and complex software, detecting commonly-used design patterns, and software evolution. In this work, the software structure is represented as a call graph depicting software artefacts and their direct dependencies. These call graphs are compared based on graph mining approach to detect similarities between two software structures. The proposed method is adapted and applied to two industrial use cases with different size and complexity. The obtained similar fragments in the software structures are evaluated and verified through manual analysis. The results show that the proposed method is promising approach to capture the similarities between two software structures.