Towards Dynamic Consistency Checking in Goal-directed Predicate Answer Set Progr
summary
Dynamic Consistency Checking interleaves generation of goal-directed partial stable models in s(CASP) with early evaluation of global constraints (called denials), so a literal is rejected before it enters the tentative model if it violates remaining denials. A prototype that checks ground literals shows up to 90x speedup on Hamiltonian-path and n‑queens benchmarks, cuts the number of generated models drastically, and outperforms an auxiliary-predicate encoding for queens; work continues to add constraint-domain reduction and propagation.
AI-written summary · about