Search

This thesis concerns the proof complexity of algebraic and semialgebraic proof systems Polynomial Calculus, Sums-of-Squares and Sherali-Adams. The most studied complexity measure for these systems is the degree of the proofs. This thesis concentrates on other possible complexity measures of interest to proof complexity, monomial-size and bit-complexity. We aim to showcase that there is a reasonably well-behaved theory for these measures also. Firstly we tie the complexity measures of degree and monomial size together by proving a size-degree trade-off for Sums-of-Squares and Sherali-Adams. We show that if there is a refutation with at most s many monomials, then there is a refutation whose degree is of order square root of n log s plus k, where k is the maximum degree of the constraints and n is the number of variables. For Polynomial Calculus similar trade-off was obtained earlier by Impagliazzo, Pudlák and Sgall. Secondly we prove a feasible interpolation property for all three systems. We show that for each system there is a polynomial time algorithm that given two sets P(x,z) and Q(y,z) of polynomial constraints in disjoint sequences x,y and z of variables, a refutation of the union of P(x,z) and Q(y,z), and an assignment a to the z-variables, finds either a refutation of P(x,a) or a refutation of Q(y,a). Finally we consider the relation between monomial-size and bit-complexity in Polynomial Calculus and Sums-of-Squares. We show that there is an unsatisfiable set of polynomial constraints that has both Polynomial Calculus and Sums-of-Squares refutations of polynomial monomial-size, but for which any Polynomial Calculus or Sums-of-Squares refutation requires exponential bit-complexity. Besides the emphasis on complexity measures other than degree, another unifying theme in all the three results is the use of semantic characterizations of resource-bounded proofs and refutations. All results make heavy use of the completeness properties of such characterizations. A, Aquesta tesi tracta de la complexitat de les proves en els sistemes de prova algebraics i semialgebraics Càlcul Polinomial (Polynomial Calculus), Sumes de Quadrats (Sums of Squares), i Sherali-Adams. La mesura de complexitat més estudiada per a aquests sistemes és el grau dels polinomis. Aquesta tesi se centra en altres possibles mesures de complexitat d'interès per a la complexitat de proves: el nombre de monomis i la longitud de representació en nombre de bits. Pretenem demostrar que aquestes mesures admeten una teoria comparable i complementària a la teoria del grau com a mesura de complexitat. En primer lloc, establim una relació entre les mesures de grau i de nombre de monomis demostrant una propietat d'intercanvi (trade-off) entre les dues mesures per als sistemes Sumes de Quadrats i Sherali-Adams. Demostrem que si hi ha una refutació amb com a màxim s monomis, aleshores hi ha una refutació el grau de la qual és d'ordre de l'arrel quadrada de n.log(s) més k, on k és el grau màxim de les restriccions i n és el nombre de variables. Per al Càlcul Polinomial, una propietat d'intercanvi similar va ser obtinguda per Impagliazzo, Pudlák i Sgall. En segon lloc, demostrem que els tres sistemes admeten la propietat d'interpolació eficient. Mostrem que, per a cadascun dels sistemes, hi ha un algorisme de temps polinomial que, donat dos conjunts P(x,z) i Q(y,z) de restriccions polinomials en successions disjuntes de variables x, y i z, donada una refutació de la unió de les restriccions de P(x,z) i Q(y,z), i donada una assignació per a les variables z, troba una refutació de P(x,a) o una refutació de Q(y,a). Finalment considerem la relació entre el nombre de monomis i la longitud de representació en bits per al Càlcul Polinomial i per a Sumes de Quadrats. Mostrem que hi ha un conjunt insatisfactible de restriccions polinomials que admet refutacions tant en Càlcul Polinomial com en Sumes de Quadrats amb un nombre polinòmic de monomis, però per a les quals qualsevol refutac, Postprint (published version)

We prove that both Polynomial Calculus and Sums-of-Squares proof systems admit a strong form of feasible interpolation property for sets of polynomial equality constraints. Precisely, given two sets P(x, z) and Q(y, z) of equality constraints, a refutation Π of P(x, z) ∪ Q(y, z), and any assignment a to the variables z, one can find a refutation of P(x, a) or a refutation of Q(y, a) in time polynomial in the length of the bit-string encoding the refutation Π. For Sums-of-Squares we rely on the use of Boolean axioms, but for Polynomial Calculus we do not assume their presence., Partially funded by European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme, grant agreement ERC-2014-CoG 648276 (AUTAR) and MICCIN grant TIN2016-76573-C2-1P (TASSAT3)., Peer Reviewed, Postprint (published version)

We show that if a system of degree-k polynomial constraints on n Boolean variables has a Sums-of-Squares (SOS) proof of unsatisfiability with at most s many monomials, then it also has one whose degree is of the order of the square root of n log s plus k. A similar statement holds for the more general Positivstellensatz (PS) proofs. This establishes size-degree trade-offs for SOS and PS that match their analogues for weaker proof systems such as Resolution, Polynomial Calculus, and the proof systems for the LP and SDP hierarchies of Lovász and Schrijver. As a corollary to this, and to the known degree lower bounds, we get optimal integrality gaps for exponential size SOS proofs for sparse random instances of the standard NP-hard constraint optimization problems. We also get exponential size SOS lower bounds for Tseitin and Knapsack formulas. The proof of our main result relies on a zero-gap duality theorem for pre-ordered vector spaces that admit an order unit, whose specialization to PS and SOS may be of independent interest., Peer Reviewed, Postprint (published version)