A Transformation Technique for Datalog Programs Based on Non-deterministic Constructs

Ποτίκας, Πέτρος; Rondogiannis, Panos; Gergatsoulis, Manolis

doi:10.1007/3-540-45607-4_2

Cited by 1 publication

(1 citation statement)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, the language Cactus [31] (and its function-free subset called Branching-time Datalog [30]) is a branching-time logic programming formalism; similarly, Datalog nS [6,7] has an extended notion of time built in its design (and has linear-time as a very special case, called Datalog 1S [4]). It should be noted that branching-time languages can express certain problems in a natural way [31] and have recently found interesting applications in the area of Datalog optimizations [30,25]. Therefore, the second issue that arises is the derivation of an efficient temporal stratification test for this more general notion of time.…”

Section: The Problemmentioning

confidence: 99%

Temporal stratification tests for linear and branching-time deductive databases

Nomikos

Rondogiannis

Gergatsoulis

2005

Theoretical Computer Science

Self Cite

View full text Add to dashboard Cite

We consider the problem of extending temporal deductive databases with stratified negation. We argue that the classical stratification test for deductive databases is too restrictive when one shifts attention to the temporal case. Moreover, as we demonstrate, the (more general) local stratification approach is impractical: detecting whether a temporal deductive database is locally stratified is shown to be co-NP hard (even if one restricts attention to programs that only use one predicate symbol and two constants). For these reasons we define temporal stratification, an intermediate notion between stratification and local stratification. We demonstrate that for the temporal deductive databases we consider, temporal stratification coincides with local stratification in certain important cases in which the latter is polynomial-time decidable. We then develop two algorithms for detecting temporal stratification. The first algorithm applies to linear-time temporal deductive databases and it is efficient and more general than existing approaches; however, the algorithm sacrifices completeness for efficiency since it does not cover the whole class of temporally stratified programs. The second algorithm applies to branching-time temporal deductive databases (which include as a special case the linear-time ones).

show abstract