MULTI-AGENTS SYSTEM
| Collection | International Encyclopedia of Systems and Cybernetics |
|---|---|
| Year | 2004 |
| Vol. (num.) | 2(2) |
| ID | ◀ 2216 ▶ |
| Object type | General information, Methodology or model |
A network composed of numerous interconnected elements or individuals.
This very general model is derived from the concept of distributed artificial intelligence.
As stated by J. ERCEAU and J. FERBER, it offers a frame for the study of: “… the behavior of a set of more or less expert entities, more or less organized along social type laws. These entities or agents benefit from a certain degree of autonomy, and are immersed in some environment in which and with which they interact. They are structured around three functions: to perceive, decide and/or act. They can be physical entities (sensors, processors, vehicles, …) or abstract ones (tasks to be performed, movements,…). Through reciprocal information transmission, they become able to act on their environment and upon themselves, i.e. to modify their own behavior. To this end, they are endowed with a partial representation of this environment and with perception and communication means”.
- “As their behavior is based on an a priori or on acquired knowledge, agents include two main features. The first one is a social tendency, oriented towards the collectivity: this is the conversational and relational domain, and the corresponding mechanisms and associated knowledge bases are related to the group's activity. The other is the individual tendency: in this case, the register is about knowledge and behavior, with mechanisms and knowledge bases including rules for the internal workings of the agent” (1991, p.752).
What the authors describe are networks of artificial agents (robots of limited individual abilities). But the model also strikingly suggests analogies with human and, probably, insects societies.
Networks of this type are most generally decentralized, since their organization, while in need and in search of coherence, is a result of more or less free, local and specific interrelations. As an example, R. BROOKS proposes micro-robots “societies” that would be used for planetary exploration, in which case the sideral distances make efficient communications with any earth based users quite difficult, due to time lags (1989).