Logic modules for communicating distributed agents

Abstract

The development of isolated agents involves the manipulation of components such as actions, mental attitudes and decisions. Multi-agent systems have to deal with these components from different agents, without to forget the privacy and autonomy of each agent. For programming simple agents, both object-oriented and logic paradigms have shown advantages in terms of behavior encapsulation and mental model manipulation, respectively. Multi-paradigm languages allowing the usage of both programming paradigms can sol ve sorne troubles in agent-oriented programming. However, problems about interaction among agents, especially in physically distributed environments, can not directly be sol ved with that solution. This paper presents an approach based on the usage of logic modules for supporting interaction among agents in a multi-paradigm environment. Logic modules encapsulate a part of a mental state of an agent, being each logic module composed by a sequence of clauses. Objects representing agents manipulate these logic modules. They can transport their own logic modules to the location of another agent. Thus, agents can integrate logic knowledge to their own decision process.