An Evaluation Framework for MAS Modeling Languages Based on Metamodel Metrics

  • Authors:

  • Iván García-Magariño;Jorge J. Gómez-Sanz;Rubén Fuentes-Fernández

  • Affiliations:

  • Dept. Software Engineering and Artificial Intelligence, Facultad de Informática, Universidad Complutense de Madrid, Spain;Dept. Software Engineering and Artificial Intelligence, Facultad de Informática, Universidad Complutense de Madrid, Spain;Dept. Software Engineering and Artificial Intelligence, Facultad de Informática, Universidad Complutense de Madrid, Spain

  • Venue:
  • Agent-Oriented Software Engineering IX
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

The fast pace of evolution in Agent-oriented Software Engineering leads to a great variety of continuously changing Multi-Agent System (MAS) Modeling Languages (MLs). In this situation, there is a rising need of evaluation for MAS MLs, as the plenty of works on this subject reflects. This paper follows this line of research presenting an evaluation framework to measure quantitatively MAS MLs. The framework includes metrics about availability, specificity, and expressiveness of the MLs. Otherwise than existing frameworks, this work considers metamodels to define its measures and focuses on the quantitative measurement instead of qualitative evaluations. With these metrics and the data gathered from existing MLs, the goal is to quantify the appropriateness of a given MAS ML for a particular problem domain. In addition, these metrics can quantitatively track the improvements of MAS MLs on these features. The paper also presents the results of the current experiments with the framework that have taken measures in nine problem domains with six MAS MLs.