Modeling with Plausibility Checking: Inspecting Favorable and Critical Signs for Consistency between Control Flow and Functional Behavior

Abstract

UML activity diagrams are a wide-spread modelling technique to capture behavioral aspects of system models. Usually, pre- and post-conditions of activities are described in natural language and are not formally integrated with the static domain model. Hence, early consistency validation of activity models is difficult due to their semiformal nature. In this paper, we use integrated behavior models that integrate activity diagrams with object rules defining sets of actions in simple activities. We formalize integrated behavior models using typed, attributed graph transformation. It provides a basis for plausibility checking by static conflict and causality detection between specific object rules, taking into account their occurrence within the control flow. This analysis leads to favorable as well as critical signs for consistency of the integrated behavior model. Our approach is supported by ActiGra, an Eclipse plug-in for editing, simulating and analyzing integrated behavior models. It visualizes favorable and critical signs for consistency in a convenient way and uses the well-known graph transformation tool AGG for rule application as well as static conflict and causality detection. We validate our approach by modeling a conference scheduling system.