Behavior Differentiation of Process Variants With Invisible Tasks

It is a fact that plenty of process variants are derived from the same base model in practical applications, and main reasons include inevitable software maintenance adaptability models. This raises question many with high similarity degree, can not be differentiated each other under state-of-the-art measurements. In order to differentiate similar-but-different variants, we propose an approach variants’ behavior differentiation this paper. The method analyzes complex structures invisible tasks, utilizes task execution relationship construct integrated measurement, which extends existing measurement capacity terms dealing tasks. proved proposed capture dominate features variant including means as long behaviors two different, corresponding matrices must different. Furthermore, set experiments carried out, evaluate properties effectiveness, semantic uniqueness expression correctness method. Meanwhile, experimental results give evidences outperforms measurements accuracy.