Matthew L. Ginsberg

TL;DR: This paper develops a search procedure that is more likely to find a solution in any given time limit than existing methods such as chronological backtracking and iterative sampling and shows formally and experimentally that limited discrepancy search can be expected to outperform existing approaches.

TL;DR: This paper describes a uniform formalization of much of the current work in artificial intelligence on inference systems, showing that many of these systems, including first‐order theorem provers, assumption‐based truth maintenance systems, and unimplemented formal systems such as default logic or circumscription, can be subsumed under a single general framework.

TL;DR: The technique developed is a variant of dependency-directed backtracking that uses only polynomial space while still providing useful control information and retaining the completeness guarantees provided by earlier approaches.

TL;DR: This work presents a general scheme whereby symmetries are exploited by adding \symmetry-breaking" predicates to the theory, and discusses methods for generating partial symmetry-breaking predicates, and shows that in several speciic cases asymmetries can be broken either fully or partially using a polynomial number of predicates.