Positive semidefinite rank

Similar documents
Semidefinite programming lifts and sparse sums-of-squares

Convex sets, conic matrix factorizations and conic rank lower bounds

Equivariant semidefinite lifts and sum of squares hierarchies

Convex sets, matrix factorizations and positive semidefinite rank

Power and limitations of convex formulations via linear and semidefinite programming lifts. Hamza Fawzi

Lifts of Convex Sets and Cone Factorizations

Lower bounds on the size of semidefinite relaxations. David Steurer Cornell

LIFTS OF CONVEX SETS AND CONE FACTORIZATIONS JOÃO GOUVEIA, PABLO A. PARRILO, AND REKHA THOMAS

Symmetric Sums of Squares

Polygons as sections of higher-dimensional polytopes

arxiv: v2 [math.mg] 10 Feb 2015

6-1 The Positivstellensatz P. Parrilo and S. Lall, ECC

LMI MODELLING 4. CONVEX LMI MODELLING. Didier HENRION. LAAS-CNRS Toulouse, FR Czech Tech Univ Prague, CZ. Universidad de Valladolid, SP March 2009

Hilbert s 17th Problem to Semidefinite Programming & Convex Algebraic Geometry

SPECTRAHEDRA. Bernd Sturmfels UC Berkeley

1 Quantum states and von Neumann entropy

The convex algebraic geometry of rank minimization

Approximate cone factorizations and lifts of polytopes

Semidefinite Programming

Semidefinite Programming

Sta$s$cal models and graphs

The maximal stable set problem : Copositive programming and Semidefinite Relaxations

Exact algorithms: from Semidefinite to Hyperbolic programming

On the existence of 0/1 polytopes with high semidefinite extension complexity

MIT Algebraic techniques and semidefinite optimization February 14, Lecture 3

Computation of the Joint Spectral Radius with Optimization Techniques

Semi-definite representibility. For fun and profit

The Graph Realization Problem

Spherical Euclidean Distance Embedding of a Graph

SEMIDEFINITE PROGRAM BASICS. Contents

A Hierarchy of Polyhedral Approximations of Robust Semidefinite Programs

Robust and Optimal Control, Spring 2015

MIT Algebraic techniques and semidefinite optimization May 9, Lecture 21. Lecturer: Pablo A. Parrilo Scribe:???

Chapter 2: Linear Programming Basics. (Bertsimas & Tsitsiklis, Chapter 1)

A notion of Total Dual Integrality for Convex, Semidefinite and Extended Formulations

Semidefinite approximations of the matrix logarithm

Dimension reduction for semidefinite programming

COURSE ON LMI PART I.2 GEOMETRY OF LMI SETS. Didier HENRION henrion

The Geometry of Semidefinite Programming. Bernd Sturmfels UC Berkeley

Convex Optimization. (EE227A: UC Berkeley) Lecture 28. Suvrit Sra. (Algebra + Optimization) 02 May, 2013

9. Interpretations, Lifting, SOS and Moments

Agenda. Applications of semidefinite programming. 1 Control and system theory. 2 Combinatorial and nonconvex optimization

CSC Linear Programming and Combinatorial Optimization Lecture 10: Semidefinite Programming

Copositive Programming and Combinatorial Optimization

The Algebraic Degree of Semidefinite Programming

ALGEBRA: From Linear to Non-Linear. Bernd Sturmfels University of California at Berkeley

3. Vector spaces 3.1 Linear dependence and independence 3.2 Basis and dimension. 5. Extreme points and basic feasible solutions

4. Algebra and Duality

Relaxations and Randomized Methods for Nonconvex QCQPs

WHICH NONNEGATIVE MATRICES ARE SLACK MATRICES?

Facial reduction for Euclidean distance matrix problems

A CONIC DANTZIG-WOLFE DECOMPOSITION APPROACH FOR LARGE SCALE SEMIDEFINITE PROGRAMMING

SDP Relaxations for MAXCUT

Semidefinite programs and combinatorial optimization

Semidefinite and Second Order Cone Programming Seminar Fall 2001 Lecture 5

Lecture 4: Polynomial Optimization

SPECTRAHEDRA. Bernd Sturmfels UC Berkeley

Complexity of Deciding Convexity in Polynomial Optimization

Rational Sums of Squares and Applications

Complexity of 10 Decision Problems in Continuous Time Dynamical Systems. Amir Ali Ahmadi IBM Watson Research Center

What can be expressed via Conic Quadratic and Semidefinite Programming?

Semidefinite programming and convex algebraic geometry

On the existence of 0/1 polytopes with high semidefinite extension complexity

Lecture 6: Conic Optimization September 8

Convex algebraic geometry, optimization and applications

A NEW SEMIDEFINITE PROGRAMMING HIERARCHY FOR CYCLES IN BINARY MATROIDS AND CUTS IN GRAPHS

Advances in Convex Optimization: Theory, Algorithms, and Applications

CONVEX ALGEBRAIC GEOMETRY. Bernd Sturmfels UC Berkeley

Canonical Problem Forms. Ryan Tibshirani Convex Optimization

Cuts for mixed 0-1 conic programs

Robust linear optimization under general norms

Copositive Programming and Combinatorial Optimization

Applications of semidefinite programming in Algebraic Combinatorics

Convex relaxation. In example below, we have N = 6, and the cut we are considering

Convex relaxation. In example below, we have N = 6, and the cut we are considering

Notes on the decomposition result of Karlin et al. [2] for the hierarchy of Lasserre by M. Laurent, December 13, 2012

Nonconvex Quadratic Programming: Return of the Boolean Quadric Polytope

Information Theory + Polyhedral Combinatorics

Representations of Positive Polynomials: Theory, Practice, and

On the local stability of semidefinite relaxations

Lecture 1. 1 Conic programming. MA 796S: Convex Optimization and Interior Point Methods October 8, Consider the conic program. min.

Acyclic Semidefinite Approximations of Quadratically Constrained Quadratic Programs

CSC Linear Programming and Combinatorial Optimization Lecture 12: The Lift and Project Method

E5295/5B5749 Convex optimization with engineering applications. Lecture 5. Convex programming and semidefinite programming

Decentralized Control of Stochastic Systems

LOVÁSZ-SCHRIJVER SDP-OPERATOR AND A SUPERCLASS OF NEAR-PERFECT GRAPHS

1 The independent set problem

A positive definite polynomial Hessian that does not factor

CONTAINMENT PROBLEMS FOR POLYTOPES AND SPECTRAHEDRA

Semidefinite Programming Basics and Applications

Agenda. 1 Cone programming. 2 Convex cones. 3 Generalized inequalities. 4 Linear programming (LP) 5 Second-order cone programming (SOCP)

Lift-and-Project Techniques and SDP Hierarchies

Lecture: Introduction to LP, SDP and SOCP

Unbounded Convex Semialgebraic Sets as Spectrahedral Shadows

Example: feasibility. Interpretation as formal proof. Example: linear inequalities and Farkas lemma

Semidefinite and Second Order Cone Programming Seminar Fall 2001 Lecture 4

Robust conic quadratic programming with ellipsoidal uncertainties

Analysis and synthesis: a complexity perspective

3. Linear Programming and Polyhedral Combinatorics

CS675: Convex and Combinatorial Optimization Fall 2016 Combinatorial Problems as Linear and Convex Programs. Instructor: Shaddin Dughmi

Transcription:

1/15 Positive semidefinite rank Hamza Fawzi (MIT, LIDS) Joint work with João Gouveia (Coimbra), Pablo Parrilo (MIT), Richard Robinson (Microsoft), James Saunderson (Monash), Rekha Thomas (UW) DIMACS Workshop on Distance Geometry July 2016

2/15 Euclidean distance matrices Theorem (Schoenberg, 1935) M is an Euclidean distance matrix if and only if diag(m) = 0 and [M 1,i + M 1,j M i,j ] 2 i,j n is positive semidefinite. Allows us to express certain distance geometry problems as semidefinite programs

2/15 Euclidean distance matrices Theorem (Schoenberg, 1935) M is an Euclidean distance matrix if and only if diag(m) = 0 and [M 1,i + M 1,j M i,j ] 2 i,j n is positive semidefinite. Allows us to express certain distance geometry problems as semidefinite programs Which convex sets can be represented using semidefinite programming?

3/15 Semidefinite representation Feasible set of a semidefinite program: { X 0 (positive semidefinite constraint) A(X ) = b (linear constraints)

3/15 Semidefinite representation Feasible set of a semidefinite program: { X 0 (positive semidefinite constraint) A(X ) = b (linear constraints) Convex set C has a semidefinite representation of size d if: C = π(s d + L) S d + = d d positive semidefinite matrices L = affine subspace π = linear map L S d + π C

4/15 Examples of semidefinite representations Examples: EDM n+1 has SDP representation of size n

4/15 Examples of semidefinite representations Examples: EDM n+1 has SDP representation of size n Disk in R 2 has a SDP representation of size 2 [ ] x 2 + y 2 1 x y 1 0 y 1 + x

4/15 Examples of semidefinite representations Examples: EDM n+1 has SDP representation of size n Disk in R 2 has a SDP representation of size 2 [ ] x 2 + y 2 1 x y 1 0 y 1 + x Square [ 1, 1] 2 has a SDP representation of size 3 [ 1, 1] 2 = (x1, x2) R2 : u R 1 x 1 x 2 x 1 1 u 0 x 2 u 1

5/15 Existential question vs. complexity question Existential question: Which convex sets admit a semidefinite representation? Helton-Nie conjecture: Any convex set defined using polynomial inequalities has a semidefinite representation

5/15 Existential question vs. complexity question Existential question: Which convex sets admit a semidefinite representation? Helton-Nie conjecture: Any convex set defined using polynomial inequalities has a semidefinite representation Complexity question: Given a convex set C, what is smallest semidefinite representation of C?

5/15 Existential question vs. complexity question Existential question: Which convex sets admit a semidefinite representation? Helton-Nie conjecture: Any convex set defined using polynomial inequalities has a semidefinite representation Complexity question: Given a convex set C, what is smallest semidefinite representation of C? Positive semidefinite rank

Importance of lifting 6/15

Importance of lifting 6/15

Importance of lifting 6/15

6/15 Importance of lifting Ben-Tal and Nemirovski: Regular polygon with 2 n sides can be described using only n inequalities!

6/15 Importance of lifting Ben-Tal and Nemirovski: Regular polygon with 2 n sides can be described using only n inequalities! Lift = inverse of elimination (cf. Pablo s talk)

7/15 Lifts of polytopes and ranks of matrices P polytope in R d Slack matrix of P: Nonnegative matrix M of size #facets(p) #vertices(p): where M i,j = h i gi T v j gi T x h i are the facet inequalities of P v j are the vertices of P v j h i g T i v j g T i x h i

7/15 Lifts of polytopes and ranks of matrices P polytope in R d Slack matrix of P: Nonnegative matrix M of size #facets(p) #vertices(p): where M i,j = h i gi T v j gi T x h i are the facet inequalities of P v j are the vertices of P v j h i g T i v j g T i x h i

7/15 Lifts of polytopes and ranks of matrices P polytope in R d Slack matrix of P: Nonnegative matrix M of size #facets(p) #vertices(p): where M i,j = h i gi T v j gi T x h i are the facet inequalities of P v j are the vertices of P v j h i g T i v j g T i x h i x 1 1 0 1 1 0 1 0 1 0

7/15 Lifts of polytopes and ranks of matrices P polytope in R d Slack matrix of P: Nonnegative matrix M of size #facets(p) #vertices(p): where M i,j = h i gi T v j gi T x h i are the facet inequalities of P v j are the vertices of P v j h i g T i v j g T i x h i 0 1 1 0 0 0 1 1 0 0 1 1

7/15 Lifts of polytopes and ranks of matrices P polytope in R d Slack matrix of P: Nonnegative matrix M of size #facets(p) #vertices(p): where M i,j = h i gi T v j gi T x h i are the facet inequalities of P v j are the vertices of P v j h i g T i v j g T i x h i 0 1 1 0 0 0 1 1 1 0 0 1 0 1 0 1

7/15 Lifts of polytopes and ranks of matrices P polytope in R d Slack matrix of P: Nonnegative matrix M of size #facets(p) #vertices(p): where M i,j = h i gi T v j gi T x h i are the facet inequalities of P v j are the vertices of P v j h i g T i v j g T i x h i 0 1 1 0 0 0 1 1 1 0 0 1 1 1 0 0 1 1 0 0

8/15 Positive semidefinite rank M R p q with nonnegative entries Positive semidefinite factorization: M ij = Tr(A i B j ), where A i, B j S k + rank psd (M) = size of smallest psd factorization B j A i Tr(A i B j )

9/15 Example Consider M ij = (i j) 2 for 1 i, j n: 0 1 4 9 16 1 0 1 4 9 M = 4 1 0 1 4 9 4 1 0 1 16 9 4 1 0 rank psd (M) = 2 (independent of n): Let A i = [ ] [ ] [ ] T 1 i 1 1 i i 2 = and B i i j = One can verify that M ij = Tr(A j B j ). [ ] j 2 j = j 1 [ j 1 ] [ ] T j. 1

10/15 SDP representations and psd rank Theorem (Gouveia, Parrilo, Thomas, 2011) Let P be polytope with slack matrix M. The smallest semidefinite representation of P has size exactly rank psd (M). Polytope P Slack matrix M rank psd (M) Works for more generally for convex sets (slack matrix is infinite) Proof based on duality for semidefinite programming

10/15 SDP representations and psd rank Theorem (Gouveia, Parrilo, Thomas, 2011) Let P be polytope with slack matrix M. The smallest semidefinite representation of P has size exactly rank psd (M). Polytope P Slack matrix M rank psd (M) Works for more generally for convex sets (slack matrix is infinite) Proof based on duality for semidefinite programming Example: Slack matrix of square [ 1, 1] 2 has positive semidefinite rank 3.

11/15 Properties of rank psd Satisfies the usual properties one would expect for a rank (invariance under scaling, subadditivity, etc.) [Fawzi, Gouveia, Parrilo, Robinson, Thomas, Positive semidefinite rank, Math. Prog., 2015] Connection with problems in information theory NP-hard to compute [Shitov, 2016]

12/15 Linear programming (LP) lifts Polytope P has LP lift of size d if it can be written as P = π(r d + L) where L affine subspace and π linear map Nonnegative factorization of M of size d: M ij = a T i b j where a i, b j R d + rank + (M) := size of smallest nonnegative factorization of M Theorem (Yannakakis, 1991) Let P be polytope with slack matrix M. The smallest LP lift of P has size exactly rank + (M).

13/15 LP lifts vs. SDP lifts Example The square P = [ 1, 1] 2 : SDP lifts: P has an SDP lift of size 3: [ 1, 1] 2 = (x1, x2) R2 : u R 1 x 1 x 2 x 1 1 u 0 x 2 u 1 SDP lift of size 3. LP lifts: Can show that any LP lift of [ 1, 1] 2 must have size 4. Stable set polytope for perfect graphs: SDP lift of linear size (Lovász) but no currently known LP lift of polynomial size

14/15 LP lifts vs. SDP lifts Question: How powerful are SDP lifts compared to LP lifts? Theorem (Fawzi, Saunderson, Parrilo, 2015) There is a family of polytopes P d R 2d such that ( ) rank psd (P d ) log d rank + (P d ) O 0. d P d = trigonometric cyclic polytope (generalization of regular polygons) Construction uses tools from Fourier analysis + sparse sums of squares

15/15 Conclusion Semidefinite representations of convex sets Connection with matrix factorization Linear programming vs. semidefinite programming lifts for polytopes For more information: [Fawzi, Gouveia, Parrilo, Robinson, Thomas, Positive semidefinite rank, Math. Prog., 2015]

15/15 Conclusion Semidefinite representations of convex sets Connection with matrix factorization Linear programming vs. semidefinite programming lifts for polytopes For more information: [Fawzi, Gouveia, Parrilo, Robinson, Thomas, Positive semidefinite rank, Math. Prog., 2015] Thank you!