Hybridizing evolutionary algorithms and clustering algorithms to find source-code clones

This paper presents a hybrid approach to detect source-code clones that combines evolutionary algorithms and clustering. A case-study is conducted on a small C++ code base. The preliminary investigation indicates that such an approach is effective in detecting groups of source-code clones. Categories and Subject Descriptors D.2.7 [Software Engineering]: Restructuring, reverse engineering, and reengineering, I.2.8 [Artificial Intelligence]: Problem Solving, Control Methods, and Search. General Terms Algorithms, Design, Experimentation. Keywords Evolutionary Algorithms, Software Engineering, Clone Detection.

• Type: Conference
• Year: 2005
• Where: GECCO
• Authors: Andrew Sutton, Huzefa H. Kagdi, Jonathan I. Maletic, L. Gwenn Volkert
• Keywords: Evolutionary Algorithms, GECCO 2005, Software Engineering, Source-code Clones