Using Answer Sets to Solve Belief Change Problems

We describe BE, an implemented system for solving belief change problems in the presence of actions. We illustrate how we can use BE to compute the result of belief progression, belief revision, and belief evolution in a transition system framework. The basic idea is to identify belief change problems with path-finding problems in a transition system. First, BE translates belief change problems into logic programs with answer sets corresponding to an agent’s evolving beliefs. Second, it uses existing smodels libraries to find the answer sets for the given logic programs. Conflicting observations are handled in a separate preprocessing phase, which is also based on finding answer sets.

• Type: Conference
• Year: 2007
• Where: LPNMR
• Authors: Aaron Hunter, James P. Delgrande, Joel Faber
• Keywords: Answer Sets, Belief Change Problems, LPNMR 2007, Transition System