A Multi-agent System Framework for Dialogue Games in the Group Decision-Making Context

Abstract

Dialogue games have been applied to various contexts in computer science and artificial intelligence, particularly to define interactions between autonomous software agents. However, in order to implement dialogue games, the developers need to deal with other important details besides what is presented in the model’s definition. This is a complex work, mostly when it is expected that the agents’ interactions correctly represent a human group behavior. In this work, we present a multi-agent system framework specifically designed to facilitate the implementation of dialogue games under the context of group decision-making in which agents interact as the humans do in face-to-face meetings. The proposed framework, named MAS4GDM, encapsulates the JADE framework and provides a layer that allows developers to easily implement their dialogue models without being concerned with some complex implementation details, such as: the communication model, the agents’ life cycle, among others. We ran an experimental evaluation and verified that the proposed framework allows to implement dialogue models in an easier way and abstract the developers from important implementation details that can compromise the application’s success.