|
FLoC
MEETINGS
PROGRAM
FACILITIES
SEATTLE
ORGANIZATION
MISCELLANEOUS
OUT-OF-DATE
|
WST on Wednesday, August 16th
| 09:30‑10:30 |
Yuri Gurevich
(Microsoft Research Redmond)
Program termination and well partial orderings [ppt]
The following known observation may be useful in establishing program termination:
if a transitive relation R is covered by finitely many well-founded relations
U1, ..., Un then R is well-founded.
A question arises how to bound the ordinal height |R| of the relation
R in terms of the ordinals αi = |Ui|.
We introduce the notion of the stature ||P|| of a well partial
ordering P and show that
|R| ≤ ||α1 * ... * αn||.
Furthermore, this bound is tight; the equality is achieved in some cases.
The notion of stature is of considerable independent interest. We
define ||P|| as the ordinal height of the forest of nonempty bad
sequences of P, but it has many other natural and equivalent
definitions. In particular, ||P|| is the supremum, and in fact the
maximum, of the lengths of linearizations of P. And
||α1 * ... * αn|| is equal to the
natural product
α1 ⊗ ... ⊗ αn.
Coming back to the question above, we have that |R| ≤
α1 ⊗ ... ⊗ αn,
and the bound is tight.
 
|
| 11:00‑11:30 |
Harald Zankl (University of Innsbruck)
Nao Hirokawa (University of Innsbruck)
Aart Middeldorp
(University of Innsbruck)
Constraints for Argument Filterings
Argument filterings are a key ingredient of the dependency pair method.
Finding a suitable argument filtering that enables the constraints
originating from the dependency pair method to be solved by a strictly
monotone base order is a challenging search problem. In this note we
propose a simple encoding of argument filterings in propositional logic
which can be easily combined with propositional encodings of
simplification orders, resulting in a propositional formula with the
property that any satisfying assignment corresponds to an argument
filtering and the parameters of the encoded order which solve the
constraints and vice-versa. In order to test the effectiveness of our
approach, we implemented the combination with the embedding order on
top of the recursive SCC algorithm of Hirokawa and Middeldorp. For
satisfiability checking we use two different methods, the
state-of-the-art SAT solver MiniSat and a simple OCaml package for
manipulating OBDDs. The results are compared with the divide and
conquer algorithm implemented in TTT.
 
|
| 11:30‑12:00 |
Harald Zankl (University of Innsbruck)
Aart Middeldorp
(University of Innsbruck)
KBO as a Satisfaction Problem
This note presents an approach to prove termination of term rewrite
systems (TRSs) with the Knuth-Bendix order efficiently. The constraints
for the weight function as well as for the precedence are encoded in
propositional logic and the resulting formula is tested for
satisfiability. Any satisfying assignment represents a weight function
and a precedence such the induced Knuth-Bendix order orients the rules
of the encoded TRS from left to right.
 
|
| 12:00‑12:30 |
Michael Codish (Ben-Gurion University)
Peter Schneider-Kamp
(RWTH Aachen)
Vitaly Lagoon (University of Melbourne)
Rene Thiemann
(RWTH Aachen)
Juergen Giesl
(RWTH Aachen)
Automating Dependency Pairs Using SAT Solvers
Recent results provide a propositional encoding for lexicographic
path orders (LPOs) which facilitates the application of SAT solvers
for termination analysis of term rewrite systems.
Using this encoding as a black-box together with
dependency pairs is not very efficient. Therefore, in this paper
we investigate a new combined encoding of LPO in connection with
dependency pairs. This new encoding solves two main issues:
the combined search for a LPO together with an
argument filtering to orient a set of inequalities, and
how the choice of the argument filtering influences the set of
inequalities that have to be oriented.
We have implemented our contributions in the termination prover AProVE.
Extensive experiments show that by our new encoding and the application of SAT
solvers one obtains speedups in orders of magnitude as well as increased
termination proving power.
 
|
| 14:00‑14:30 |
Matt Kaufmann
(University of Texas)
Panagiotis Manolios
(Georgia Tech)
J Strother Moore
(University of Texas at Austin)
Daron Vroon
(Georgia Institute of Technology, College of Computing)
Integrating CCG Analysis into ACL2
ACL2 is a powerful, industrial strength theorem proving system, which has been used on
numerous verification pro jects. It is part of the Boyer-Moore family of provers, for which its authors received the 2005 ACM Software System Award. Termination plays a central role in ACL2’s logic. It is used to demonstrate the logical consistency of function definitions and allows for the admission of an induction scheme that mirrors each function’s recursive structure.
In previous work we introduced calling context graphs (CCGs) and showed how they can be combined with theorem proving queries to provide a powerful method for proving termination of programs written in first order, functional programming languages. In this paper we describe work we are doing to integrate CCG-based termination analysis into ACL2.
 
|
| 14:30‑15:00 |
Frederic Blanqui
(INRIA)
Adam Koprowski
(Eindhoven University of Technology)
Solange Coupet-Grimal
(Université de Provence, Aix-Marseille I)
William Delobel (Université de Provence, Aix-Marseille I)
Sebastien Hinderer
(LORIA)
CoLoR, a Coq library on Rewriting and termination
Coq is a tool allowing to certify proofs. This paper describes a Coq
library for certifying termination proofs.
 
|
| 15:00‑15:30 |
Johannes Waldmann
(HTWK Leipzig, FB IMN)
Free SCC Analysis via Constant Interpretations
We show how to split a (relative)
top termination problem into independent subproblems,
by using interpretations that are constant on strict rules.
This can typically be applied to top termination problems
that arise from the dependency pair transformation.
The method is applied in the termination prover Matchbox.
 
|
|
|