Modeling Self-* Systems Using Chemically-Inspired Composable Patterns

Abstract

The behaviour of self-* systems is complex to model from an algorithmic point of view. Designing and specifying self-* systems implies a great amount of work that can be sensibly reduced if models can be reused and composed in a modular way. This article discusses a chemically-inspired architecture and formalisms that facilitate the creation of modular, reusable models based on behavioural patterns inspired by behaviours found in nature. The architecture is based on chemical-like laws ruling the evolution of the system. We show the reuse of general behavioural patterns using three concrete examples of self-* systems from different domains.