Extracting Proofs from Branch-and-Prune

نویسندگان

  • Sicun Gao
  • Soonho Kong
  • Michael Wang
  • Edmund M. Clarke
چکیده

δ-Complete decision procedures can solve SMT problems over the reals with a wide range of nonlinear functions, allowing “δ-bounded errors”. The scalability of such procedures usually depends on efficient numerical procedures, whose implementation can be error-prone. It is important for δ-complete solvers to provide certificates to prove the correctness of their answers. We show how to do this for DPLL〈ICP〉, a general solving framework based on Interval Constraint Propagation. We focus on the construction of proof trees for the “unsat” answers and the proof-checking of their correctness. Besides certifying solvers, we find our approach a promising one for automated theorem proving over the reals, exploiting the power of numerical algorithms in a formal way. One direct application is to establish many nonlinear lemmas in the Flyspeck project, for the formal proof of the Kepler Conjecture. This research was sponsored by the National Science Foundation grants no. DMS1068829, no. CNS0926181 and no. CNS0931985, the GSRC under contract no. 1041377, the Semiconductor Research Corporation under contract no. 2005TJ1366, and the Office of Naval Research under award no. N000141010188.

برای دانلود رایگان متن کامل این مقاله و بیش از 32 میلیون مقاله دیگر ابتدا ثبت نام کنید

ثبت نام

اگر عضو سایت هستید لطفا وارد حساب کاربری خود شوید

منابع مشابه

A Branch-and-Prune Algorithm for the Molecular Distance Geometry Problem

The Molecular Distance Geometry Problem consists in finding the positions in R of the atoms of a molecule, given some of the inter-atomic distances. We show that under an additional requirement on the given distances (which is realistic from the chemical point of view) this can be transformed to a combinatorial problem. We propose a Branch-and-Prune algorithm for the solution of this problem an...

متن کامل

Simplifying and Generalizing Formulae in Tableaux. Pruning the Search Space and Building Models

A powerful extension of the tableau method is described. It consists in a new simpliication rule allowing to prune the search space and a new way of extracting a model from a given (possibly innnite) branch. These features are combined with a former method for simultaneous search for refutations and models. The possibilities of the new method w.r.t. the original one are clearly stated. In parti...

متن کامل

Excluding Regions Using Sobol Sequences in an Interval Branch-and-Prune Method for Nonlinear Systems

Traditional rejection/reduction tests used in branch-and-prune methods for nonlinear systems usually are based on various forms of the interval Newton operator and constraint propagation techniques. Hence, they are relatively costly. This paper considers an additional phase of a branchand-prune method for the exclusion of regions not containing any solutions. Low-discrepancy sequences of Sobol ...

متن کامل

MD-jeep: An Implementation of a Branch and Prune Algorithm for Distance Geometry Problems

We present MD-jeep, an implementation of a Branch & Prune (BP) algorithm, which we employ for the solution of distance geometry problems related to molecular conformations. We consider the problem of finding the conformation of a molecule from the distances between some pairs of its atoms, which can be estimated by experimental techniques. We reformulate this problem as a combinatorial optimiza...

متن کامل

A Symbolic-Numerical Branch and Prune Algorithm for Solving Non-linear Polynomial Systems

This paper discusses the processing of non-linear polynomial systems using a branch and prune algorithm within the framework of constraint programming. We propose a formalism for a kind of branch and prune algorithm implementing symbolic and numerical methods to reduce the systems with respect to a relation de ned from both inclusion of variable domains and inclusion of sets of constraints. The...

متن کامل

ذخیره در منابع من


  با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید

عنوان ژورنال:

دوره   شماره 

صفحات  -

تاریخ انتشار 2013