Lazy-Grounding for Answer-Set Programming: Challenges and Potential

Antonius Weinzierl

Answer-Set Programming (ASP) is a well-known formalism for logic-based AI where knowledge is expressed by first-order rules and solutions to problems are given in terms of answer sets. ASP features non-monotonic reasoning and is widely used in a variety of fields, thanks to efficient solving techniques that are heavily inspired by SAT solving. Traditional ASP solving, however, is also hampered by the so-called grounding bottleneck, which often consumes all available memory, because first-order rules are grounded upfront and grounding may have to consider all possible variable substitutions instead of just the really relevant ones. Lazy grounding is a technique that avoids the grounding bottleneck, but it also introduces new obstacles for efficient solving. This talk introduces lazy grounding for ASP solving, presents challenges and available solutions, as well as an outlook at potential for new solving techniques.`