Using ranking functions to determine plausible action histories

Hunter, Aaron (Aaron Hunter (Aaron_Hunter)) (author)
Delgrande, James, P. (author)
(editor)
(translator)

© 2005 Aaron Hunter and James P. Delgrande

Proceedings of the Sixth Workshop on Nonmonotonic Reasoning, Action, and Change (NRAC-05), Edinburgh, Scotland, 2005. We use ranking functions to reason about belief change following an alternating sequence of actions and observations. At each instant, an agent assigns a plausibility value to every action and every state; the most plausible world histories are obtained by minimizing the sum of these values. Since plausibility is given a quantitative rank, an agent is able to compare the plausibility of actions and observations. This allows action occurrences to be postulated or refuted in response to new observations. We demonstrate that our formalism is a generalization of our previous work on the interaction of revision and update.

http://www.cse.unsw.edu.au/~nrac05/Proceedings/Papers/nrac05-proceedings...