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.

378 CHAPTER 5. EUCLIDEAN DISTANCE MATRIX<br />

find<br />

G ℵ ∈S M , G X ∈S N , Z∈R M×N G X<br />

subject to d i ≤ tr(G X Φ i ) ≤ d i , ∀i ∈ I 1<br />

cos ϕ j ≤ tr(G ℵ B j ) ≤ cos ϕ j , ∀j ∈ I 2<br />

〈Z , A k 〉 = 0 , ∀k ∈ I 3<br />

G X 1 = 0<br />

δ(G ℵ ) = 1<br />

[ ]<br />

Gℵ Z<br />

Z T ≽ 0<br />

G X<br />

[ ]<br />

Gℵ Z<br />

rank<br />

Z T = 3<br />

G X<br />

(857)<br />

where G X 1=0 provides a geometrically centered list X (5.4.2.2). Ignoring<br />

the rank constraint would tend to force cross-product matrix Z to zero.<br />

What binds these three variables is the rank constraint; we show how to<br />

satisfy it in4.4.<br />

<br />

5.4.3 Inner-product form EDM definition<br />

[p.20] We might, for example, realize a constellation given only<br />

interstellar distance (or, equivalently, distance from Earth and<br />

relative angular measurement; the Earth as vertex to two stars).<br />

Equivalent to (794) is [332,1-7] [287,3.2]<br />

d ij = d ik + d kj − 2 √ d ik d kj cos θ ikj<br />

= [√ d ik<br />

√<br />

dkj<br />

] [ 1 −e ıθ ikj<br />

−e −ıθ ikj<br />

1<br />

] [√ ]<br />

d ik<br />

√<br />

dkj<br />

(858)<br />

called the law of cosines, where ı ∆ = √ −1 , i,k,j are positive integers, and<br />

θ ikj is the angle at vertex x k formed by vectors x i − x k and x j − x k ;<br />

cos θ ikj =<br />

1<br />

(d 2 ik + d kj − d ij )<br />

√ = (x i − x k ) T (x j − x k )<br />

dik d kj ‖x i − x k ‖ ‖x j − x k ‖<br />

(859)

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

Saved successfully!

Ooh no, something went wrong!