Matching UML class diagrams using a Hybridized Greedy-Genetic algorithm

Abstract

Model matching is a fundamental operation for various model management aspects such as model retrieval, evolution, and merging. An accurate matching between the elements of the matched models results in a better model management. This paper presents a Hybridized Greedy-Genetic algorithm for matching UML class diagrams, considering their lexical, internal, and structural similarity. Additionally, using a case study of five class diagrams, the performance of the Hybridized algorithm is empirically compared against the traditional Genetic algorithm in terms of both matching accuracy and convergence time.