2001, Volume 4, Number 3, pp.212-249

This is the second part of an overview devoted to autonomous agents and multi-agent systems. Such systems are deemed as designed to obtain an ensemble of interconnected agents that can operate beyond the capabilities of any individual. Cooperation and competition among interacting agents are regarded as a foundation of self-organization of the system. Collective intelligence is viewed as emergent property that can arise in a system of agents due to self-organization under a certain level of complexity. Neural networks and neural fields are considered as dynamical systems that can be used for implementation of intelligent agents. In this connection some feasible dynamical mechanisms of thinking are considered which lead to an architecture of cognitive neural systems and intelligent agents. It is also discussed a cognitive model in which cognition is treated as anticipation through creating and matching hypotheses, which might appear essential for intelligent agent and brain building.
Key words: autonomous agents, multi-agent systems, neural networks, neural fields, nonlinear dynamical systems, self-organization, artificial intelligence, thinking, cognition, anticipation, language, meanings of information, brain building

Full text:  Acrobat PDF  (302KB)   PostScript (1460KB)   PostScript.gz (303KB)

Copyright © Nonlinear Phenomena in Complex Systems. Last updated: October 15, 2001