Modelling of Multi-Agent Systems: Experiences with Membrane Computing and Future Challenges

TL;DR

This paper reviews a decade of research on formal modelling of Multi-Agent Systems using membrane computing, highlighting challenges, solutions, and future directions for dynamic, complex, and emergent systems.

Contribution

It presents novel applications of membrane computing, specifically P Systems, for modeling dynamic and complex MAS like social insects and swarms.

Findings

Membrane computing effectively models dynamic MAS structures.

P Systems capture emergent behaviors in biological-inspired systems.

Research identifies key challenges and solutions in MAS formalization.

Abstract

Formal modelling of Multi-Agent Systems (MAS) is a challenging task due to high complexity, interaction, parallelism and continuous change of roles and organisation between agents. In this paper we record our research experience on formal modelling of MAS. We review our research throughout the last decade, by describing the problems we have encountered and the decisions we have made towards resolving them and providing solutions. Much of this work involved membrane computing and classes of P Systems, such as Tissue and Population P Systems, targeted to the modelling of MAS whose dynamic structure is a prominent characteristic. More particularly, social insects (such as colonies of ants, bees, etc.), biology inspired swarms and systems with emergent behaviour are indicative examples for which we developed formal MAS models. Here, we aim to review our work and disseminate our findings to fellow researchers who might face similar challenges and, furthermore, to discuss important issues for advancing research on the application of membrane computing in MAS modelling.