A generative approach for multi-agent system development

  • Authors:

  • Uirá Kulesza;Alessandro Garcia;Carlos Lucena;Paulo Alencar

  • Affiliations:

  • PUC-Rio, Computer Science Department, LES, SoC+Agents Group, Rio de Janeiro, RJ, Brazil;PUC-Rio, Computer Science Department, LES, SoC+Agents Group, Rio de Janeiro, RJ, Brazil;PUC-Rio, Computer Science Department, LES, SoC+Agents Group, Rio de Janeiro, RJ, Brazil;Computer Science Department, Computer System Group, University of Waterloo, Waterloo, Ontario, Canada

  • Venue:
  • Software Engineering for Multi-Agent Systems III
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The development of Multi-Agent Systems (MASs) involves special concerns, such as interaction, adaptation, autonomy, among others. Many of these concerns are overlapping, crosscut each other and the agent’s basic functionality. Over the last few years, several methodologies and implementation frameworks have been proposed to support agent-oriented software engineering. Although these approaches have brought some benefits to improve the productivity and quality on the MAS development, they present some restrictions. First, agent-oriented methodologies are too high level and do not indicate how to master the complexity of MAS concerns based on the object-oriented abstractions. Second, implementation frameworks provide object-oriented APIs for MAS development without providing guidelines for the modularization of agent concerns. Moreover, neither of the proposed agent oriented-approaches deals with the modeling and implementation of agent crosscutting concerns. This paper presents a generative approach for the development of MASs that addresses these restrictions. The proposed approach explores the MAS domain to enable the code generation of heterogeneous agent architectures. Aspect-oriented techniques are used to allow the modeling of crosscutting agent features. The generative approach brings several benefits to the code generation and modeling of agent crosscutting features since early development stages.