Constraints among argument sizes in logic programs (extended abstract)
Kirack Sohn · 1994
In logic programs the argument sizes of derivable facts w.r.t. an n-ary predicate are viewed as a set of points in Rn, which are approximated by their convex hull. Interargument constraint w.r.t. a predicate is essentially a set of constraints that every derivable fact of the predicate satisfies. We formalize such constraints by a fixpoint of recursive transformation similar to immediate consequence operator. However, the transformation does not necessarily converge finitely. Approximating polycones to their affine hulls provides useful interargument constraints in many practical programs, guaranteeing finite convergence. For a class of linear recursive logic programs satisfying translativeness property, precise interargument constraints can be obtained by an analysis of structures of recursive transformations.