International Journal of Agent-Oriented Software Engineering 2(2), pages 222–245
2008
Systems Biology promotes a system-level understanding of biological systems, and requires modelling and simulating tools for understanding, controlling and re-creating biological systems and their dynamics. The articulation of multiagent systems (MAS) in terms of multiple, distributed and autonomous computational entities makes MAS a seemingly fit paradigm for modelling and simulating biological systems and networks according to the System Biology perspective.
In this paper we adopt the A&A (agents and artifacts) meta-model – where the notions of agent, artifact, and workspace are taken as the basic bricks for MAS – as the ontological foundation for our multi-agent-based simulation (MABS) framework, and discuss how this impacts on the modelling and simulation of biological systems. After re-casting the A&A abstractions within the domain and design models, we specialise A&A within the System Biology context, and show a possible operational model based on the TuCSoN agent coordination infrastructure, upon which our simulation framework is implemented. There, agents – representing active biological components such as proteins – interact by means of artifacts built upon TuCSoN tuple centres – representing the bio-chemical environment that enables, mediates and govern the interaction of biological components – within workspaces—representing different spatial regions, like cell compartments. As a case study, we model and simulate a well-studied metabolic pathway such as glycolysis, and present some results of the simulation.
keywords
Systems biology; multiagent-based simulation; MABS; A&A metamodelling; MAS engineering; TuCSoN; multi-agent systems; agent-based systems; agents and artefacts; system dynamics; metabolic pathway; glycolysis
journal or series
International Journal of Agent-Oriented Software Engineering
(IJAOSE)
