Objectif
Phase transitions in combinatorial optimization is a new, exciting research direction at the crossroads of statistical mechanics, combinatorial optimization and artificial intelligence.
Its goals are to use intuitions from statistical physics to shed light on the underlying reasons for computational intractability, thus complementing computational complexity theory. Despite significant progress, phase transitions are still largely a case-by-case approach, with few connections to computational complexity theory.
We propose:
- to develop the phase transition approach into a systematic theory, related to Computational Complexity, using insights and methods from Computational Complexity theory; and
- to increase the practicality of this approach, by investigating phase transitions for instances with a regular structure, with a special focus on those arising from problems in formal verification.
A benefit of this approach (if successful) would be bringing the algorithmic advances experienced in the area of satisfiability solving by the definition of the survey propagation algorithm to the area of formal verification.
Champ scientifique
Appel à propositions
FP6-2004-MOBILITY-12
Voir d’autres projets de cet appel