Book Reviews ARCHON: An Architecture for Multi-agent Systems. Wittig, Thies (Ed.) West Sussex, UK: Ellis Horwood Limited (distributed by Prentice Hall). 1992, 135 pp. Munindar P. Singh Department of Computer Science North Carolina State University their Archon layers--work like humans might in a similar system. Thus a cooperative system is produced from isolated components. It is assumed that the underlying component is a knowledgebased system. The authors argue that this is not restrictive. However, they also state that the Archon layer is kept generic by making the underlying systems be more than just knowledge-based. The underlying systems must be intelligent and reflective so as to give rationales for their decisions and to infer what information is missing or erroneous. This appears to be a strong assumption, because most expert systems would not support it. This assumption is not strongly justified within the book. What is clear and novel about the architecture is the role of cooperation among agents. The agents autonomously detect the need to cooperate--this generalizes distributed problem solving, and enhances the autonomy of the agents. The agents maintain self models and acquaintance models to effectively decide when and how to cooperate. Decision making about cooperation and
/lp/association-for-computing-machinery/book-review-archon-an-architecture-for-multi-agent-systems-edited-by-0DO0b0Sse2