Towards a domain-specific modeling language for self-adaptive systems conceptual modeling

Abstract

Self-adaptive Systems (SaSs) are able to adapt their behavior at runtime in response to contextual changes. In this work, we are interested in SaSs conceptual modeling, which is the act of creating models that describe aspects of the world. SaSs modeling is a non-trivial activity because it deals with requirements uncertainty, contextual changes, and behavior adaptation. This complexity can be minimized by using Domain-Specific Modeling Languages (DSMLs), which may be created by extending Unified Modeling Language (UML). In this paper, we propose a UML profile that represents the higher-level abstractions required to provide support for SaSs conceptual modeling. We developed the UML profile by modeling the domain of interest and extending the UML class metaclass. The UML profile was evaluated through the focus group technique, which was performed by software engineering professors. As the outcome, the focus group participants considered the UML profile able to produce SaSs conceptual models with more expressiveness than UML standard.