Tim Kluge

Title: Context-adaptive coworking with robots and roles

Abstract:

In recent years, software has continued to spread into all aspects of life. An increasing amount of cyber-physical systems directly interact with the real world, e. g. autonomous drones. These systems are expected to adapt to changing circumstances autonomously while collaborating to achieve common goals more efficiently. Such systems are known as collective adaptive systems (CAS).

While prior work in the RoSI research training group has shown that adaptations can be modelled with roles and structured compartments, modelling the behaviour of the adaptation, i. e. when to adapt and thus when to collaborate in a CAS, has not been investigated yet.

To address this, we propose to model adaptations using rewriting rules on the role-play graph defined in compartments. Complete adaptability of the system is provided by meta-adaptation of the rules. Roles are played by asynchronously communicating objects following the actor model of concurrency.

In this talk, central aspects of CAS are discussed. Based on the identified features, we compare our solution to the state of the art. Two case studies are presented to evaluate our approach.