Characterizing causal action theories and their implementations in answer set programming

Mapping Action Language BC to Logic Programs: A Characterization by Postulates

“…We then show that these four properties provide a unique characterization of the BC semantics under the so-called permissible mappings. We then relate our result to those in (Zhang and Lin 2015) for action languages B and C, and conclude the paper.…”

“…Recently Zhang and Lin (2015) proposed a set of properties for characterizing action languages B (Gelfond and Lifschitz 1998) and C (Giunchiglia and Lifschitz 1998). They considered static causal rules of the form (l, G), where l is a fluent literal and G a set of fluent literals.…”

“…While these rules can be interpreted by either B's semantics or C's semantics, they cannot be mixed. This is reflected in incompatible sets of postulates that we used for capturing B and C in (Zhang and Lin 2015). However, this presents a problem if one tries to capture the language BC (Lee, Lifschitz, and Yang 2013).…”

“…In terms of logic programs, such a rule is translated as l ← G, not not H or l ← G, not H using strong negation. Thus the semantics of BC is a combination of B and C: G is translated according to B while H according to C. However, given that a static causal rule in (Zhang and Lin 2015) is a pair (l, G), the semantics of BC cannot be captured in their framework.…”

“…To address this problem, we propose to extend static causal rules in (Zhang and Lin 2015) to triples (l, G, H), with G as premises and H justifications. We'll show with this extension, the mixed semantics of BC on static causal rules can be similarly characterized.…”