Abstract
Agent architectures provide the blueprints for the design and development of individual agents. The purpose of agent architecture is to define modes and modules of agent’s interaction in the world as well as connections among internal components of an agent. An agent can be understood in terms of its architectural description of its perceptual, deliberation, and acting capabilities. Architectures have been used to describe robot control software [Musliner, Durfee, and Shin, 1993]. Such architectures emphasized rapid processing of early perception and low-level actuation capabilities. Brooks’ subsumption architecture has been an influential guide for fast and reactive robotic actions [Brook 1986]. Although subsumption was good for real-time processing needed in robotics, it never became useful for agents. This is partly because being reactive is a standard property of agents. Pro-active architectures such as logic-based, utility-based, or belief-desire-intention (BDI) have been more popular in agent architectures [Wooldridge, 2000]. Intentional agents are modeled in multi-modal BDI logics. Each architecture has its strengths and weaknesses that make it suitable for particular roles or particular types of problems. Instead of comparing and contrasting architectures, here we give a partial list of collective properties for pro-active architectures: logical correctness, ability to initiate a new course of action, ability to form and manipulate explicit intentions, ability to reason based on nontrivial models of the world, ability to adapt, ability to learn, ability to interact with emotions, and ability to react when there is not enough time to complete reasoning. In nontrivial monolithic systems, proactive architectures addressed many issues including: world modeling, modularity of cognitive functions (such as planning and learning), affect, and uncertainty management. For agent-based systems, reasoning about autonomy is a specific area of concern. Finally, nontrivial agents who have to account for other agents and be social must address many issues. A partial list is: coordination, cooperation, teamwork, and other relationships and social attitudes such as autonomy, veracity, awareness, benevolence, rationality, roles, obligations.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Brooks, R.: A Robust Layered Control Systems for a Mobile Robot. IEEE Journal of Robotics and Automation RA 2-1, 14–23 (1986)
Musliner, D., Durfee, E., Shin, K.: CIRCA: A Cooperative Intelligent Real-Time Control Architecture. IEEE Transactions on Systems, Man, and Cybernetics (1993)
Wooldridge, M.: Reasoning about Rational Agents. The MIT Press, Cambridge (2000)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2003 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Hexmoor, H. (2003). Evolution of Agent Architectures. In: Truszkowski, W., Hinchey, M., Rouff, C. (eds) Innovative Concepts for Agent-Based Systems. WRAC 2002. Lecture Notes in Computer Science(), vol 2564. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-45173-0_40
Download citation
DOI: https://doi.org/10.1007/978-3-540-45173-0_40
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-40725-6
Online ISBN: 978-3-540-45173-0
eBook Packages: Springer Book Archive