Rule-based modeling and static analysis of self-adaptive systems by graph transformation

Abstract

Software systems nowadays require continuous operation despite changes both in user needs and in their operational environments. Self-adaptive systems are typically instrumented with tools to autonomously perform adaptation to these changes while maintaining some desired properties. In this paper we model and analyze self-adaptive systems by means of typed, attributed graph grammars. The interplay of different grammars representing the application and the adaptation logic is realized by an adaption manager. Within this formal framework we define consistency and operational properties that are maintained despite adaptations and we give static conditions for their verification. The overall approach is supported by the AGG tool for modeling, simulating, and analyzing graph transformation systems. A case study modeling a business process that adapts to changing environment conditions is used to demonstrate and validate the formal framework.