Propositional circumscription and extended closed-world reasoning are _ 2^ P-complete

From MaRDI portal
Publication:2367539





For non-monotonic logics (logic programming, default reasoning, etc.), several versions of Closed World Assumption are known. All of them express the idea that, crudely speaking, all possible knowledge about the domain area is already included in the knowledge base, so if some statement is not deducible from the knowledge base, then it must be false. The authors analyze the complexity of the following deduction problem: given a (finite) propositional theory with an appropriate version of CWA, and a query \(Q\), to check whether \(T\) (together with this version of CWA) implies \(Q\). For all known versions of CWA, except the so-called ``simple CWA, this deduction problem is proven to be \(\Pi^ P_ 2\)-hard (i.e., at least as hard as the problem of checking whether a formula of the type \(\forall \vec x\;\exists \vec y\;B( \vec x,\vec y,\vec u)\) is satisfiable, where \(\vec x= (x_ 1,\dots,x_ n)\), \(\vec y=(y_ 1,\dots,y_ m)\), and \(\vec u=(u_ 1,\dots,u_ l)\) are sets of Boolean variables). For simple CWA, the deduction problem belongs to the class \(\Delta^ P_ 2\), and, most probably, is even easier, but it is still harder than the deduction problem in normal propositional logic (without CWA).



Cites work


Cited in
(50)








This page was built for publication: Propositional circumscription and extended closed-world reasoning are \(\Pi_ 2^ P\)-complete

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2367539)