Facial Reduction for Symmetry Reduced Semidefinite Programs

We consider both facial and symmetry reduction techniques for semidefinite programming, SDP. We show that the two together fit surprisingly well in an alternating direction method of multipliers, ADMM, approach. The combination of facial and symmetry reduction leads to a significant improvement in both numerical stability and running time for both the ADMM and interior point approaches. We test our method on various doubly nonnegative, DNN, relaxations of hard combinatorial problems including quadratic assignment problems with sizes of more than $n=500$.

[1]  J. Lofberg,et al.  YALMIP : a toolbox for modeling and optimization in MATLAB , 2004, 2004 IEEE International Conference on Robotics and Automation (IEEE Cat. No.04CH37508).

[2]  Frank Permenter,et al.  Reduction methods in semidefinite and conic optimization , 2017 .

[3]  Alexander J. Smola,et al.  Graph Partitioning via Parallel Submodular Approximation to Accelerate Distributed Machine Learning , 2015, ArXiv.

[4]  Franz Rendl,et al.  A Copositive Programming Approach to Graph Partitioning , 2007, SIAM J. Optim..

[5]  Franz Rendl,et al.  Semidefinite Programming Relaxations for the Quadratic Assignment Problem , 1998, J. Comb. Optim..

[6]  Renata Sotirov,et al.  An Efficient Semidefinite Programming Relaxation for the Graph Partition Problem , 2014, INFORMS J. Comput..

[7]  J. Borwein.  Characterization of optimality for the abstract convex program with finite dimensional range , 1981, Journal of the Australian Mathematical Society.

[8]  Jos F. Sturm,et al.  A Matlab toolbox for optimization over symmetric cones , 1999 .

[9]  Burak Eksioglu,et al.  Clustering of high throughput gene expression data , 2012, Comput. Oper. Res..

[10]  Frank Thomson Leighton,et al.  A Framework for Solving VLSI Graph Layout Problems , 1983, J. Comput. Syst. Sci..

[11]  J. Borwein,et al.  Regularizing the Abstract Convex Program , 1981 .

[12]  Etienne de Klerk,et al.  Exploiting group symmetry in semidefinite programming relaxations of the quadratic assignment problem , 2007, Math. Program..

[13]  Renata Sotirov,et al.  Semidefinite programming and eigenvalue bounds for the graph partition problem , 2013, Math. Program..

[14]  Jiming Peng,et al.  Estimating Bounds for Quadratic Assignment Problems Associated with Hamming and Manhattan Distance Matrices Based on Semidefinite Programming , 2010, SIAM J. Optim..

[15]  Dion Gijswijt,et al.  Matrix Algebras and Semidefinite Programming Techniques for Codes , 2005, 1007.0906.

[16]  Franz Rendl,et al.  Copositive and semidefinite relaxations of the quadratic assignment problem , 2009, Discret. Optim..

[17]  Knud D. Andersen,et al.  The Mosek Interior Point Optimizer for Linear Programming: An Implementation of the Homogeneous Algorithm , 2000 .

[18]  Laurent Condat.  Fast projection onto the simplex and the l1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\pmb {l}_\mathbf {1}$$\end{ , 2015, Mathematical Programming.

[19]  Alexander Schrijver,et al.  Invariant Semidefinite Programs , 2010, 1007.2905.

[20]  Alexander Schrijver,et al.  A comparison of the Delsarte and Lovász bounds , 1979, IEEE Trans. Inf. Theory.

[21]  Qing Zhao,et al.  Semidefinite Programming Relaxations for the Graph Partitioning Problem , 1999, Discret. Appl. Math..

[22]  Dmitriy Drusvyatskiy,et al.  The Many Faces of Degeneracy in Conic Optimization , 2017, Found. Trends Optim..

[23]  Etienne de Klerk,et al.  Numerical block diagonalization of matrix *-algebras with application to semidefinite programming , 2011, Math. Program..

[24]  Franz Rendl,et al.  The min-cut and vertex separator problem , 2018, Comput. Optim. Appl..

[25]  Johan Efberg,et al.  YALMIP : A toolbox for modeling and optimization in MATLAB , 2004 .

[26]  Etienne de Klerk,et al.  Improved semidefinite programming bounds for quadratic assignment problems with suitable symmetry , 2012, Math. Program..

[27]  Franz Rendl,et al.  QAPLIB – A Quadratic Assignment Problem Library , 1997, J. Glob. Optim..

[28]  Henry Wolkowicz,et al.  ADMM for the SDP relaxation of the QAP , 2015, Math. Program. Comput..

[29]  Hao Sun,et al.  A strictly contractive Peaceman-Rachford splitting method for the doubly nonnegative relaxation of the minimum cut problem , 2021, Computational Optimization and Applications.

[30]  Franz Rendl,et al.  A spectral approach to bandwidth and separator problems in graphs , 1993, IPCO.

[31]  Bin Fu,et al.  Multi-directional Width-Bounded Geometric Separator and Protein Folding , 2005, ISAAC.

[32]  Hao Sun,et al.  Eigenvalue, quadratic programming, and semidefinite programming relaxations for a cut minimization problem , 2016, Comput. Optim. Appl..

[33]  Laurent Condat,et al.  A Fast Projection onto the Simplex and the l 1 Ball , 2015 .

[34]  F. Vallentin.  Symmetry in semidefinite programs , 2007, 0706.4233.

[35]  U. Truetsch.  A semidefinite programming based branch-and-bound framework for the quadratic assignment problem , 2014 .

[36]  Richard F. Barrett,et al.  Matrix Market: a web resource for test matrix collections , 1996, Quality of Numerical Software.

[37]  P. Parrilo,et al.  Symmetry groups, semidefinite programs, and sums of squares , 2002, math/0211450.

[38]  Robert E. Tarjan,et al.  Applications of a planar separator theorem , 1977, 18th Annual Symposium on Foundations of Computer Science (sfcs 1977).

[39]  Renata Sotirov,et al.  Graph bisection revisited , 2016, Annals of Operations Research.

[40]  Yunmei Chen,et al.  Projection Onto A Simplex , 2011, 1101.6081.

[41]  J. Borwein,et al.  Facial reduction for a cone-convex programming problem , 1981, Journal of the Australian Mathematical Society. Series A. Pure Mathematics and Statistics.

[42]  D. Rose,et al.  Generalized nested dissection , 1977 .

[43]  Stephen P. Boyd,et al.  Distributed Optimization and Statistical Learning via the Alternating Direction Method of Multipliers , 2011, Found. Trends Mach. Learn..

[44]  Henry Wolkowicz,et al.  Error Bounds and Singularity Degree in Semidefinite Programming , 2019, SIAM J. Optim..

[45]  Hao Hu,et al.  On Solving the Quadratic Shortest Path Problem , 2017, INFORMS J. Comput..

[46]  Etienne de Klerk,et al.  Exploiting special structure in semidefinite programming: A survey of theory and applications , 2010, Eur. J. Oper. Res..

[47]  P. Delsarte.  AN ALGEBRAIC APPROACH TO THE ASSOCIATION SCHEMES OF CODING THEORY , 2011 .

[48]  J. Wedderburn,et al.  On Hypercomplex Numbers , 1908 .