On the stability of (self-)adaptive behaviour in continuously changing environments: A quantification approach

Abstract

The concept of self-adaptation and self-organisation (SASO) is a modern approach to cope with the ever-increasing complexity and interconnectedness of large-scale component systems. The basic idea is to react to environmental dynamics and disturbances by re-configuring the productive behaviour and/or the relations to other systems. However, this may result in unstable and even oscillating macro-level behaviour, potentially rendering the adaptation efforts of the contained component systems inappropriate. We assume that such an unstable configuration is an indicator of unexpected behaviour which can lead to a reduced utility of the overall system. To enable the system to be self-aware about such events, we propose a concept to measure the configuration stability of a SASO system by creating a derived time series based on the configurations. This is based on the application of the Kinoshita measure. We show the applicability of the concept and the observed behaviour in different simulated use-cases.

Keywords

• Self-awareness
• Self-adaptive and self-organising systems
• System analysis
• Stability
• Degree of self-adaptation
• Autonomic computing