International Encyclopedia of Systems and Cybernetics

An active element in a multi-elements system or network.

J. ERCEAU and J. FERBER describe the following types of agents, at different hierarchical levels in the active multi-agents system:

- "reactive agents: these are at the lower levels. They merely dispose of a reduced protocole and communication language and … their abilities rely only on a stimulus/action rule. The reactive agents class include various levels, according to their group-forming ability and capacity to produce global behavior;

- "communicating agents, which possess a complete communication protocole, but whose conversational and behavioral parts are interdependent;

- "rational agents, which possess precise abilities, beliefs and a partial representation of their environment, specially of the other agents within the system;

- "intentional agents, at the highest level, possessing explicit goals, specific plans which allow them to fulfill their goals, as well as the possibility to commit themselves to specific tasks, that they are obliged to carry out, or to contract other agents to execute certain actions". (1991 , p. 757-8)

This could be a stimulating description for a model of any society.

More recently, J. FERBER (1999) has given a much more precise definition of an agent:

Agent is a virtual or physical entity which:

1) is capable of acting in an environment

2) can communicate directly with other agents

3) is driven by a set of tendencies (in the form of individual objectives or of a satisfaction/survival function which it tries to optimize)

4) possesses resources of its own

5) is capable to perceive its environment (but up to a limited extent)

6) has only a partial representation of this environment (and perhaps none at all)

7) possesses skills and can offer services

8) may be able to reproduce itself

9) whose behavior tends towards satisfying its objectives, taking account of the resources and skills available to it and depending on its perception, its representation and the communication it receives

"Having the properties 1-9) an agent can be considered as an "intelligent system"(lbid)

From this description N. SAFFARPOUR (2000,p. 75) deduces the following characteristics of agents

"- Agents are autonomous, i.e. have control over their own actions

- Agents contain some level of intelligence, from fixed rule to learning engine that allows them to adapt to change in the environment

- Agents don't only act reactively, but sometimes also proactively and don't simply act in response to environment, in other words agents are goal oriented

- Agents have social ability, that is they communicate with the user, the system and other agents as required

- Agents may also cooperate with other agents to carry out more complex tasks that those they themselves can handle

- Agents may move from one system to another to access remote recourse or even to meet other agents

- Agents are adaptive, that is change their behavior based on previous experience

All these specifications are quite significant in a general sense, even if some terms used may seem ambiguous (as f. ex. "intelligence", "move from one system to another")

→ Adaptability; Artificial life; Autonomy; Behavior (Anticipatory); Intelligence (Distributed artificial); Stigmergy; Swarm