Exact Algorithms for Semidefinite Programs with Degenerate Feasible Set

Let A0, ..., An be m x m symmetric matrices with entries in Q, and let A(x) be the linear pencil A0+x1 A1 + ··· + xn An, where x=(x1,...,xn) are unknowns. The linear matrix inequality (LMI) A(x) ≥ 0 defines the subset of Rn, called spectrahedron, containing all points x such that A(x) has non-negative eigenvalues. The minimization of linear functions over spectrahedra is called semidefinite programming (SDP). Such problems appear frequently in control theory and real algebra, especially in the context of nonnegativity certificates for multivariate polynomials based on sums of squares. Numerical software for solving SDP are mostly based on the interior point method, assuming some non-degeneracy properties such as the existence of interior points in the admissible set. In this paper, we design an exact algorithm based on symbolic homotopy for solving semidefinite programs without assumptions on the feasible set, and we analyze its complexity. Because of the exactness of the output, it cannot compete with numerical routines in practice but we prove that solving such problems can be done in polynomial time if either n or m is fixed.

[1]  S. Basu,et al.  A New Algorithm to Find a Point in Every Cell Defined by a Family of Polynomials , 1998 .

[2]  Vladimir A. Yakubovich,et al.  Linear Matrix Inequalities in System and Control Theory (S. Boyd, L. E. Ghaoui, E. Feron, and V. Balakrishnan) , 1995, SIAM Rev..

[3]  Mohab Safey El Din,et al.  Real root finding for determinants of linear matrices , 2014, J. Symb. Comput..

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

[5]  Simone Naldi.  Solving Rank-Constrained Semidefinite Programs in Exact Arithmetic , 2016, ISSAC.

[6]  Pablo A. Parrilo,et al.  Semidefinite programming relaxations for semialgebraic problems , 2003, Math. Program..

[7]  Éric Schost,et al.  Bit complexity for multi-homogeneous polynomial system solving - Application to polynomial minimization , 2016, J. Symb. Comput..

[8]  Palaiseau Cedex,et al.  Computing Parametric Geometric Resolutions , 2001 .

[9]  Simone Naldi,et al.  Solving Rank-Constrained Semidefinite Programs in Exact Arithmetic , 2016, J. Symb. Comput..

[10]  Rekha R. Thomas,et al.  Semidefinite Optimization and Convex Algebraic Geometry , 2012 .

[11]  Bernd Sturmfels,et al.  The algebraic degree of semidefinite programming , 2010, Math. Program..

[12]  J. Lasserre,et al.  Handbook on Semidefinite, Conic and Polynomial Optimization , 2012 .

[13]  Marc Giusti,et al.  Intrinsic complexity estimates in polynomial optimization , 2013, J. Complex..

[14]  Mohab Safey El Din,et al.  SPECTRA – a Maple library for solving linear matrix inequalities in exact arithmetic , 2016, Optim. Methods Softw..

[15]  Éric Schost,et al.  A Nearly Optimal Algorithm for Deciding Connectivity Queries in Smooth and Bounded Real Algebraic Sets , 2013, J. ACM.

[16]  Xavier Allamigeon,et al.  Solving Generic Nonarchimedean Semidefinite Programs Using Stochastic Game Algorithms , 2016, ISSAC.

[17]  Michel Demazure,et al.  Bifurcations and Catastrophes: Geometry Of Solutions To Nonlinear Problems , 2000 .

[18]  Jean B. Lasserre,et al.  Global Optimization with Polynomials and the Problem of Moments , 2000, SIAM J. Optim..

[19]  Grigoriy Blekherman,et al.  Nonnegative Polynomials and Sums of Squares , 2010, 1010.3465.

[20]  Yurii Nesterov,et al.  Interior-point polynomial algorithms in convex programming , 1994, Siam studies in applied mathematics.

[21]  Kim-Chuan Toh,et al.  SDPT3 -- A Matlab Software Package for Semidefinite Programming , 1996 .

[22]  Mohab Safey El Din,et al.  Exact algorithms for linear matrix inequalities , 2015, SIAM J. Optim..

[23]  Etienne de Klerk,et al.  On the Convergence of the Central Path in Semidefinite Optimization , 2002, SIAM J. Optim..

[24]  Mohab Safey El Din,et al.  Testing Sign Conditions on a Multivariate Polynomial and Applications , 2007, Math. Comput. Sci..

[25]  Marie-Françoise Roy,et al.  Real algebraic geometry , 1992 .

[26]  Fabrice Rouillier,et al.  Finding at Least One Point in Each Connected Component of a Real Algebraic Set Defined by a Single Equation , 2000, J. Complex..

[27]  Éric Schost,et al.  Polar varieties and computation of one point in each connected component of a smooth real algebraic set , 2003, ISSAC '03.

[28]  S. Basu,et al.  Algorithms in real algebraic geometry , 2003 .

[29]  Motakuri V. Ramana,et al.  An exact duality theory for semidefinite programming and its complexity implications , 1997, Math. Program..

[30]  Lihong Zhi,et al.  Computing Rational Points in Convex Semialgebraic Sets and Sum of Squares Decompositions , 2010, SIAM J. Optim..

[31]  Jean-Michel Muller,et al.  Implementation and Performance Evaluation of an Extended Precision Floating-Point Arithmetic Library for High-Accuracy Semidefinite Programming , 2017, 2017 IEEE 24th Symposium on Computer Arithmetic (ARITH).

[32]  Xavier Allamigeon,et al.  Long and winding central paths , 2014, 1405.4161.

[33]  E. Yaz.  Linear Matrix Inequalities In System And Control Theory , 1998, Proceedings of the IEEE.

[34]  Lihong Zhi,et al.  Computing rational solutions of linear matrix inequalities , 2013, ISSAC '13.

[35]  Claus Scheiderer,et al.  Sums of squares of polynomials with rational coefficients , 2012, 1209.2976.

[36]  Stephen P. Boyd,et al.  Linear Matrix Inequalities in Systems and Control Theory , 1994 .

[37]  Jean B. Lasserre,et al.  Introduction to Semidefinite, Conic and Polynomial Optimization , 2012 .

[38]  Hoon Hong,et al.  Improving root separation bounds , 2018, J. Symb. Comput..

[39]  Marc Giusti,et al.  A Gröbner Free Alternative for Polynomial System Solving , 2001, J. Complex..