Probablistic Bigraphs

TL;DR

This paper introduces probabilistic and action bigraphs, extending the bigraph formalism to model stochastic, non-deterministic, and reward-based systems, with implementations in the BigraphER toolkit demonstrated through practical examples.

Contribution

It extends bigraphs to probabilistic and action variants, implements these in BigraphER, and provides the first direct implementation of the membrane budding model.

Findings

Implemented probabilistic and action bigraphs in BigraphER.

Demonstrated applications in virus spread and data harvesting.

Supported stochastic bigraphs extension and membrane budding model.

Abstract

Bigraphs are a universal computational modelling formalism for the spatial and temporal evolution of a system in which entities can be added and removed. We extend bigraphs to probablistic bigraphs, and then again to action bigraphs, which include non-determinism and rewards. The extensions are implemented in the BigraphER toolkit and illustrated through examples of virus spread in computer networks and data harvesting in wireless sensor systems. BigraphER also supports the existing stochastic bigraphs extension of Krivine et al., and using BigraphER we give, for the first time, a direct implementation of the membrane budding model used to motivate stochastic bigraphs.