Towards an Adaptive and Normative Multi-Agent System Metamodel and Language: Existing Approaches and Research Opportunities

TL;DR

This paper surveys existing Multi-Agent System (MAS) metamodels and languages, highlighting gaps in norm and adaptation support, and proposes a new MAS metamodel to address these limitations.

Contribution

It compares current MAS models' expressiveness regarding norms and adaptation, and introduces a novel metamodel to enhance support for these features.

Findings

Current MAS models lack comprehensive norm and adaptation support

Comparison reveals gaps in expressiveness of existing models

Proposes a new MAS metamodel to improve abstraction capabilities

Abstract

Several Multi-Agent System (MAS) metamodels and languages have been proposed in the literature to support the development of agent-based applications. MAS metamodels are used to capture a collection of concepts the relevant entities and relationships in the MAS domain, which include entities such as agent, message, role, action and plan, and relationships that represent, for example, that a role is responsible for one or more tasks. In addition, to models, MAS modeling languages have also been used to support the development of MASs in a wide variety of domains, including social networking, robotics, security and smart city environments. However, there is a lack of support in these models and languages for abstractions involving norms and adaptations as well as their interactions. This paper presents a survey of some existing metamodels and languages and compares their expressiveness using abstractions related to agents, norms and adaptation. The comparison serves as a basis for the definition of a new MAS metamodeling.