Title
Cones of Matrices and Successive Convex Relaxations of Nonconvex Sets
Abstract
Let F be a compact subset of the n-dimensional Euclidean space Rn represented by (finitely or infinitely many) quadratic inequalities. We propose two methods, one based on successive semidefinite programming (SDP) relaxations and the other on successive linear programming (LP) relaxations. Each of our methods generates a sequence of compact convex subsets Ck (k = 1, 2, . . .) of Rn such that (a) the convex hull of $F \subseteq C_{k+1} \subseteq C_k$ (monotonicity), (b) $\cap_{k=1}^{\infty} C_k = \text{the convex hull of F (asymptotic convergence). Our methods are extensions of the corresponding Lovász--Schrijver lift-and-project procedures with the use of SDP or LP relaxation applied to general quadratic optimization problems (QOPs) with infinitely many quadratic inequality constraints. Utilizing descriptions of sets based on cones of matrices and their duals, we establish the exact equivalence of the SDP relaxation and the semi-infinite convex QOP relaxation proposed originally by Fujie and Kojima. Using this equivalence, we investigate some fundamental features of the two methods including (a) and (b) above.
Year
DOI
Venue
2000
10.1137/S1052623498336450
SIAM Journal on Optimization
Keywords
Field
DocType
convex hull,sdp relaxation,linear ma- trix inequality,nonconvex quadratic optimization problem,nonconvex sets,successive convex relaxations,semidefinite programming,lp relaxation,quadratic inequality,exact equivalence,bilinear matrix inequality,general quadratic optimization problem,global optimization,compact convex,semi-infinite programming,compact subset,semi-infinite convex qop relaxation,quadratic inequality constraint,linear program,euclidean space,quadratic optimization,linear matrix inequality
Monotonic function,Discrete mathematics,Mathematical optimization,Combinatorics,Semi-infinite programming,Convex hull,Quadratic programming,Linear programming relaxation,Successive linear programming,Linear matrix inequality,Mathematics,Semidefinite programming
Journal
Volume
Issue
ISSN
10
3
1052-6234
Citations 
PageRank 
References 
35
5.81
12
Authors
2
Name
Order
Citations
PageRank
Masakazu Kojima11603222.51
Levent Tunçel242972.12