Artificial Bee Colony Algorithm for Lean Software Reuse

Reuse in software development is a practice that improves the software development process. Reusing existing software artifacts requires an efficient retrieval mechanism. Leaning out software repository for effective and efficient retrieval and reuse of relevant artifacts requires identifying and eliminating the wasteful artifacts it may involve. Model matching is a preliminary step to identify what is common and what is variant among the software artifacts. However, the time for matching two models to find the optimal correspondence between their elements is exponential. Artificial Bee Colony algorithm is a heuristic that is getting popularity as reasonable solution for problems under different optimization scenarios. This paper presents a solution algorithm based on Artificial Bee Colony for matching UML class diagrams. On a dataset of ten pairs of class diagrams, the performance of the suggested approach is empirically evaluated and compared with Ant Colony approach. The performance of the two algorithms are reported in terms of accuracy of matching and execution time. The results indicate the superiority of the Artificial Bee Colony algorithm in terms of accuracy rate and execution time.