DocumentCode :
3527033
Title :
A Lagrangian relaxation view of linear and semidefinite hierarchies
Author :
Lasserre, Jean B.
Author_Institution :
Inst. of Math., Univ. of Toulouse, Toulouse, France
fYear :
2013
fDate :
10-13 Dec. 2013
Firstpage :
1966
Lastpage :
1970
Abstract :
Consider the polynomial optimization problem P: f* = min{f(x): x ∈ K} where K is a compact basic semi-algebraic set. We first show that the standard Lagrangian relaxation yields a lower bound as close as desired to the global optimum f*, provided that it is applied to a problem P̃ equivalent to P, in which sufficiently many redundant constraints (products of the initial ones) are added to the initial description of P. Next we show that the standard hierarchy of LP-relaxations of P (in the spirit of Sherali-Adams´ RLT) can be interpreted as a brute force simplification of the above Lagrangian relaxation. So we provide a systematic improvement of the LP-hierarchy by doing a much less brutal simplification which results into a parametrized hierarchy of semidefinite programs (and not linear programs any more). For each semidefinite program in the hierarchy parametrized by k, the semidefinite constraint has a fixed size O(nk), independently of the rank in the hierarchy, in contrast with the standard hierarchy of semidefinite relaxations.
Keywords :
mathematical programming; polynomial approximation; relaxation theory; Lagrangian relaxation; compact basic semialgebraic set; linear hierarchy; polynomial optimization problem; semidefinite constraint; semidefinite hierarchy; semidefinite program; semidefinite relaxation; Convergence; Linear programming; Optimization; Polynomials; Programming; Standards; Vectors;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Decision and Control (CDC), 2013 IEEE 52nd Annual Conference on
Conference_Location :
Firenze
ISSN :
0743-1546
Print_ISBN :
978-1-4673-5714-2
Type :
conf
DOI :
10.1109/CDC.2013.6760169
Filename :
6760169
Link To Document :
بازگشت