"Weak" vs "Strong" Agents
"Intelligent
Agents: Theory and Practice"
Wooldridge and Jennings
http://www.doc.mmu.ac.uk/STAFF/mike/ker95.ps
- Weak agents make no commitments except to interactions
- Can use wrappers with legacy code
- Interaction model facilitates system analysis
- Much knowledge implicit in wrapper
- Coordination model is fixed
- Strong agents commit to internal structure; e.g., BDI agents
- Beliefs, Desires, Intentions
- More explicit, higher-level knowledge
- Capable of dynamic cooperation strategy choice
- Does the emphasis on and the design of internal agent structure
derive from agent functions? Really?
- What can be done with a minimal commitment to
a model of interaction?
cjp