The cut-norm ||A||C of a real matrix A=(aij)i∈ R,j∈S is the maximum, over all I ⊂ R, J ⊂ S of the quantity | Σi ∈ I, j ∈ J aij|. This concept plays a major role in the design of efficient approximation algorithms for dense graph and matrix problems. Here we show that the problem of approximating the cut-norm of a given real matrix is MAX SNP hard, and provide an efficient approximation algorithm. This algorithm finds, for a given matrix A=(aij)i ∈ R, j ∈ S, two subsets I ⊂ R and J ⊂ S, such that | Σi ∈ I, j ∈ J aij| ≥ ρ ||A||C, where ρ > 0 is an absolute constant satisfying $ρ > 0. 56. The algorithm combines semidefinite programming with a rounding technique based on Grothendieck's Inequality. We present three known proofs of Grothendieck's inequality, with the necessary modifications which emphasize their algorithmic aspects. These proofs contain rounding techniques which go beyond the random hyperplane rounding of Goemans and Williamson [12], allowing us to transfer various algorithms for dense graph and matrix problems to the sparse case.

/pdf/approximating-the-cut-norm-via-grothendieck-s-inequality-4bkj2815wu.pdf

Approximating the cut-norm via Grothendieck's inequality

This paper considers the following type of quadratic programming problem. Given an arbitrary matrix A, whose diagonal elements are zero, find x /spl isin/ {-1, 1}/sup n/ such that x/sup T/Ax is maximized. Our approximation algorithm for this problem uses the canonical semidefinite relaxation and returns a solution whose ratio to the optimum is in /spl Omega/(1/ logn). This quadratic programming problem can be seen as an extension to that of maximizing x/sup T/Ay (where y's components are also /spl plusmn/1). Grothendieck's inequality states that the ratio of the optimum value of the latter problem to the optimum of its canonical semidefinite relaxation is bounded below by a constant. The study of this type of quadratic program arose from a desire to approximate the maximum correlation in correlation clustering. Nothing substantive was known about this problem; we present an /spl Omega/ (1/logn) approximation, based on our quadratic programming algorithm. We can also guarantee that our quadratic programming algorithm returns a solution to the MAXCUT problem that has a significant advantage over a random assignment.

/pdf/maximizing-quadratic-programs-extending-grothendieck-s-w9qs471tf1.pdf

Maximizing quadratic programs: extending Grothendieck's inequality

The cut-norm $||A||_C$ of a real matrix $A=(a_{ij})_{i\in R,j\in S}$ is the maximum, over all $I \subset R$, $J \subset S$, of the quantity $|\sum_{i \in I, j\in J} a_{ij}|$. This concept plays a major role in the design of efficient approximation algorithms for dense graph and matrix problems. Here we show that the problem of approximating the cut-norm of a given real matrix is MAX SNP hard, and we provide an efficient approximation algorithm. This algorithm finds, for a given matrix $A=(a_{ij})_{i\in R,j\in S}$, two subsets $I \subset R$ and $J \subset S$, such that $|\sum_{i \in I, j\in J} a_{ij}| \geq \rho ||A||_C$, where $\rho>0$ is an absolute constant satisfying $\rho >0.56$. The algorithm combines semidefinite programming with a rounding technique based on Grothendieck's inequality. We present three known proofs of Grothendieck's inequality, with the necessary modifications which emphasize their algorithmic aspects. These proofs contain rounding techniques which go beyond the random hyperplane rounding of Goemans and Williamson [J. ACM, 42 (1995), pp. 1115-1145], allowing us to transfer various algorithms for dense graph and matrix problems to the sparse case.

Approximating the Cut-Norm via Grothendieck's Inequality

Probably the most famous of Grothendieck’s contributions to Banach space theory is the result that he himself described as “the fundamental theorem in the metric theory of tensor products”. That is now commonly referred to as “Grothendieck’s theorem” (GT in short), or sometimes as “Grothendieck’s inequality”. This had a major impact first in Banach space theory (roughly after 1968), then, later on, in C ¤ -algebra theory, (roughly after 1978). More recently, in this millennium, a new version of GT has been successfully developed in the framework of “operator spaces” or non-commutative Banach spaces. In addition, GT independently surfaced in several quite unrelated fields: in connection with Bell’s inequality in quantum mechanics, in graph theory where the Grothendieck constant of a graph has been introduced and in computer science where the Grothendieck inequality is invoked to replace certain NP hard problems by others that can be treated by “semidefinite programming’ and hence solved in polynomial time. In this expository paper, we present a review of all these topics, starting from the original GT. We concentrate on the more recent developments and merely outline those of the first Banach space period since detailed accounts of that are already available, for instance the author’s 1986 CBMS notes.

/pdf/grothendieck-s-theorem-past-and-present-1wekbrrlwf.pdf

Grothendieck's Theorem, past and present

In this paper we study semidefinite programming (SDP) models for a class of discrete and continuous quadratic optimization problems in the complex Hermitian form. These problems capture a class of well–known combinatorial optimization problems, as well as problems in control theory. For instance, they include Max–3–Cut where the Laplacian matrix is positive semidefinite (in particular, some of the edge weights can be negative). We present a generic algorithm and a unified analysis of the SDP relaxations which allow us to obtain good approximation guarantees for our models. Specifically, we give an $(k sin(\frac{\pi}{k}))^{2}/(4\pi)$ –approximation algorithm for the discrete problem where the decision variables are k–ary and the objective matrix is positive semidefinite. To the best of our knowledge, this is the first known approximation result for this family of problems. For the continuous problem where the objective matrix is positive semidefinite, we obtain the well–known π/4 result due to [2], and independently, [12]. However, our techniques simplify their analyses and provide a unified framework for treating these problems. In addition, we show for the first time that the integrality gap of the SDP relaxation is precisely π/4. We also show that the unified analysis can be used to obtain an O(1/log n)–approximation algorithm for the continuous problem in which the objective matrix is not positive semidefinite.

/pdf/on-approximating-complex-quadratic-optimization-problems-via-3k4pb5yfei.pdf

On approximating complex quadratic optimization problems via semidefinite programming relaxations

The fundamental constant of Grothendieck's inequality, defined below, was shown by Grothendieck to be less than sinh π/2=2.301+. We improve the bound slightly, and show that for the positive definite case π/2 suffices.

Ronald E. Rietz

Papers

A proof of the Grothendieck inequality