10.03.2015 Views

v2009.01.01 - Convex Optimization

v2009.01.01 - Convex Optimization

v2009.01.01 - Convex Optimization

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

256 CHAPTER 4. SEMIDEFINITE PROGRAMMING<br />

Semidefinite Farkas’ lemma provides the conditions required for a set<br />

of hyperplanes to have a nonempty intersection A ∩ S n + with the positive<br />

semidefinite cone. While the lemma as stated is correct, Ye points out<br />

[342,1.3.8] that a positive definite version of this lemma is required for<br />

semidefinite programming because any feasible point in the relative interior<br />

A ∩ int S n + is required by Slater’s condition 4.7 to achieve 0 duality gap<br />

(primal−dual objective difference4.2.3, Figure 51). In our circumstance,<br />

assuming a nonempty intersection, a positive definite lemma is required<br />

to insure a point of intersection closest to the origin is not at infinity;<br />

e.g., Figure 39. Then given A∈ R m×n(n+1)/2 having rank m , we wish to<br />

detect existence of a nonempty relative interior of the primal feasible set; 4.8<br />

b ∈ int K ⇔ 〈y, b〉 > 0 ∀y ∈ K ∗ , y ≠ 0 ⇔ A∩int S n + ≠ ∅ (599)<br />

A positive definite Farkas’ lemma can easily be constructed from this<br />

membership relation (294) and these proper convex cones K (337) and<br />

K ∗ (343):<br />

4.2.1.1.2 Lemma. Positive definite Farkas’ lemma.<br />

Given a linearly independent set {A i ∈ S n , i=1... m}<br />

b = [b i ]∈ R m , define the affine subset<br />

and a vector<br />

A = {X ∈ S n | 〈A i , X〉 = b i , i=1... m} (587)<br />

Primal feasible set relative interior A ∩ int S n + is nonempty if and only if<br />

∑<br />

y T b > 0 holds for each and every vector y = [y i ]≠ 0 such that m y i A i ≽ 0.<br />

Equivalently, primal feasible set relative interior A ∩ int S n + is nonempty<br />

if and only if y T b > 0 holds for each and every norm-1 vector ‖y‖= 1 such<br />

∑<br />

that m y i A i ≽ 0.<br />

⋄<br />

i=1<br />

4.7 Slater’s sufficient condition is satisfied whenever any primal strictly feasible point<br />

exists; id est, any point feasible with the affine equality (or affine inequality) constraint<br />

functions and relatively interior to convex cone K . [286,6.6] [36, p.325] If cone K were<br />

polyhedral, then Slater’s condition is satisfied when any feasible point exists relatively<br />

interior to K or on its relative boundary. [53,5.2.3]<br />

4.8 Detection of A ∩ int S n + by examining K interior is a trick need not be lost.<br />

i=1

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!