Mining the role-oriented process models based on genetic algorithm

Abstract

Traditional role-oriented process modeling seems to be subjective in identifying roles. To solve the problem, the similarity of activities is used in this paper. Sub-processes with high similarity are recognized as the process undertaken by a certain role. In this way, a relatively objective role identification approach is proposed, which determines the interaction between roles and establishes the role-activity diagram. Furthermore, by analyzing the interaction between roles, genetic algorithm is used to introduce multiple factors to optimize the identification. Therefore, an optimized role-oriented process modeling approach is established and an example is presented to show the feasibility of this approach. © 2012 Springer-Verlag.