A Model-Based Algorithm for Propositional Belief Revision
Fernando Zacarías, Guillermo De Ita · IEEE Latin America Transactions · 2015
Propositional Inference and belief revision are quite relevant to Automatic Reasoning. Given a KB Σ in DF and new information Σ' in CF, we show a deterministic and complete linear-time algorithm to decide Σ ⊢ Σ'. We adapt the previous algorithm to build a model-based proposal for belief revision: Σ' = Σ · P. Our proposal is based on Dalal's method of building a new DF Σ' consistent with p and whose models have the property to hold minimum changes with models of the original KB Σ. We show that, in the worst case, our proposal of belief revision involves the solution of satisfiability instances formed by subformulas of p, which implies the solution of NP-complete problems.