Learning Context Conditions for BDI Plan Selection

An important drawback to the popular Belief, Desire, and Intentions (BDI) paradigm is that such systems include no element oflearning from experience. In particular, the so-called context conditions of plans, on which the whole model relies for plan selection, are restricted to be boolean formulas that are to be specified atdesign/implementation time. To address these limitations, we propose a novel BDI programming framework that, by suitably modeling context conditions as decision trees, allows agents to learn theprobability of success for plans based on previous execution experiences. By using a probabilistic plan selection function, the agentscan balance exploration and exploitation of their plans. We developand empirically investigate two extreme approaches to learning thenew context conditions and show that both can be advantageousin certain situations. Finally, we propose a generalization of theprobabilistic plan selection function that yields a middle-groundbetween the two extreme approaches, and which we thus argue isthe most flexible and simple approach.

Authors

• Dhirendra Singh
• Sebastian Sardina
• Lin Padgham RMIT University; Patrick Lambrix
• Stephane Airiau

Keywords

• BDI
• Learning (single and multi-agent)