Control-theory and models at runtime
Pierre-Alain Muller 1, Olivier Barais 2
1 Université de Haute-Alsace
Mulhouse, France
pierre-alain.muller@uha.fr
2
IRISA / INRIA Rennes
Rennes, France
Olivier.barrais@irisa.fr
Abstract. Models at runtime are considered a key enabling technology for
systems that control themselves as they operate. The automatic-control
community has developed extensive theories and experiences in qualifying the
properties of controller and systems; including stability, observability and
commandability. In this position paper we propose to use control-theory for
describing self-adaptive model-driven systems
1 Introduction
We are witnessing significant interest in the model-driven community for using
models at runtime, beyond the now traditional construction phases of software
applications. Behind models at runtime, there is the idea that automatic or even self-
adaptability of the running system may be achieved by taking decisions based on
monitoring information captured by runtime models. In the summary of last year’s
edition of the models@runtine workshop [1], workshop organizers illustrated this by
saying that they foresaw that models could help moving a system from a consistent
architecture to another one.
While this trend is new in the modeling community, models at runtime have been
around for decades in the automatic-control community; where they have been
applied to many different domains, including: avionics, automotive, robotics, finance
and biology. There are certainly major differences behind what these two
communities recognize as models; however we have the feeling that there are also
significant commonalities, such as dealing with the behavior of dynamic systems, or
being an interdisciplinary branch of engineering and mathematics; and that the model-
driven community could benefit from part of the experience and knowledge gathered
by the automatic-control community.
2
In this short position paper we suggest that control-theory may