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.

178 CHAPTER 2. CONVEX GEOMETRY<br />

x 2<br />

1<br />

0.5<br />

0<br />

K M<br />

−0.5<br />

−1<br />

K M<br />

−1.5<br />

K ∗ M<br />

−2<br />

−1 −0.5 0 0.5 1 1.5 2<br />

x 1<br />

Figure 57: Monotone cone K M and its dual K ∗ M (drawn truncated) in R2 .<br />

Because X † x≽0 connotes membership of x to pointed K M+ , then by<br />

(270) the dual cone we seek comprises all y for which (387) holds; thus its<br />

halfspace-description<br />

K ∗ M+ = {y ≽<br />

K ∗ M+<br />

0} = {y | ∑ k<br />

i=1 y i ≥ 0, k = 1... n} = {y | X T y ≽ 0} ⊂ R n<br />

(389)<br />

The monotone nonnegative cone and its dual are simplicial, illustrated for<br />

two Euclidean spaces in Figure 56.<br />

From2.13.6.1, the extreme directions of proper K M+ are respectively<br />

orthogonal to the facets of K ∗ M+ . Because K∗ M+ is simplicial, the<br />

inward-normals to its facets constitute the linearly independent rows of X T<br />

by (389). Hence the vertex-description for K M+ employs the columns of X<br />

in agreement with Cone Table S because X † =X −1 . Likewise, the extreme<br />

directions of proper K ∗ M+ are respectively orthogonal to the facets of K M+<br />

whose inward-normals are contained in the rows of X † by (383). So the<br />

vertex-description for K ∗ M+ employs the columns of X†T . <br />

2.13.9.4.3 Example. Monotone cone.<br />

(Figure 57, Figure 58) Of nonempty interior but not pointed, the monotone<br />

cone is polyhedral and defined by the halfspace-description<br />

K M ∆ = {x ∈ R n | x 1 ≥ x 2 ≥ · · · ≥ x n } = {x ∈ R n | X ∗T x ≽ 0} (390)

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

Saved successfully!

Ooh no, something went wrong!