Intelligent Instantiation and Supersafe Rules (2016)
In the input languages of most answer set solvers, a rule with variables has, conceptually, infinitely many instances. The primary role of the process of intelligent instantiation is to identify a finite set of ground instances of rules of the given program that are "essential" for generating its stable models. This process can be launched only when all rules of the program are safe. If a program contains arithmetic operations or comparisons then its rules are expected to satisfy conditions that are even stronger than safety. This paper is an attempt to make the idea of an essential instance and the need for "supersafety" in the process of intelligent instantiation mathematically precise.
In Technical Communications of the 32nd International Conference on Logic Programming 2016.

Vladimir Lifschitz Faculty vl [at] cs utexas edu