4. Agents and multi-agent systems

Distributed problem solving and, more generally, distributed AI is a long-standing trend. The idea of distributed problem solving dates back to the mid-1970s, with actor languages on the one hand, and the blackboard architecture model on the other, initially proposed for automatic speech understanding and subsequently widely adopted. In a distributed AI system, a set of autonomous entities, called agents, interact to carry out a task contributing to the resolution of a complex problem. The agent concept has been the subject of study not only in AI, but also in philosophy and psychology. With the current development of the Internet, we are witnessing the emergence of numerous names, both French and English: resource agent, broker agent, personal assistant, mobile agent, web agent, interface agent, chatterbot, softbot, knowbot, avatar, etc.