The BDI model is well accepted as an architecture for representing and realizing rational agents. The beliefs in this model are focused on the representation of beliefs about the world and other agents and are widely independent from the agents intentions. We argue that also the representation of know-how, which captures the beliefs about actions and procedures, has to be taken into account when modeling rational agents. Using the notion of know-how as introduced by Singh we formalize and implement a concrete and usable agent architecture that supports and benefits from this representation of procedural beliefs in multiple ways. It also supports the representation of motivations that influence the agent’s behavior. We thus enable the agent to reason about its planning capabilities in the same way as it can reason about any other of its beliefs by extending a BDI-based agent architecture to allow the representation of procedural beliefs explicitly as part of the agent’s logical beliefs which again influences and enhances the agent’s behavior.
[1]
N. James,et al.
Know how
,
2002,
Antiquity.
[2]
S. Griffis.
EDITOR
,
1997,
Journal of Navigation.
[3]
Leonid Sheremetov,et al.
Weiss, Gerhard. Multiagent Systems a Modern Approach to Distributed Artificial Intelligence
,
2009
.
[4]
Julie A. Adams,et al.
Multiagent Systems: A Modern Approach to Distributed Artificial Intelligence
,
2001,
AI Mag..
[5]
Michael Luck,et al.
Motivated Behaviour for Goal Adoption
,
1998
.
[6]
Gabriele Kern-Isberner,et al.
Belief operations for motivated BDI agents
,
2008,
AAMAS.
[7]
Michael Luck,et al.
Motivated Behavior for Goal Adoption
,
1998,
DAI.
[8]
João Leite,et al.
A Survey of Programming Languages and Platforms for Multi-Agent Systems
,
2006,
Informatica.
[9]
Vladimir Lifschitz,et al.
Answer set programming and plan generation
,
2002,
Artif. Intell..
[10]
Matthias Thimm,et al.
Know-how for motivated BDI agents
,
2009,
AAMAS.