BdiIntroduction

Much theoretical work has occured that explores the consequences of different decisions and offer frameworks for reasoning about formal abstract models of mental states, there is still a large gap between this work and implemented BDI architectures.

System Complexity

BDI agent architectures tend to be very complex systems, just in terms of the amount of functionality that is built in to the architecture itself. But BDI systems appear to have significant advantages in complex multi­agent applications characterised by uncertainty, uncontrolled environmental change, and ongoing changes in specification.

BDI Advantages

• Explicit representations of the B's, D's, and I's that you can point to and reason about.
• The agents can be viewed at a level of abstraction that facilitates understanding of what it is doing and why.
• It is possible to meaningfully make statements like "Agent X wrongly believes that the fuel manifold is leaking".
• Specifications of desired system behaviour are much more easily developed into implementations of that behaviour, because the BDI architecture provides a rich, high­level "virtual machine" for the programmer that operates at a level of abstraction much closer to the specification.

• The level of abstraction for OOP is much lower than AOP and most of the design needs to be redeveloped for each application.
• We hope AOP will be easy to design and develop agents that behave in the intended manner, robustly and predictably, are easily modified, extended and controlled, and whose design can be re­used..

		[BeliefDesireIntention | BeliefDesireIntention | IntentAgents]