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.

392 CHAPTER 5. EUCLIDEAN DISTANCE MATRIX<br />

injective on domain S N h because S N⊥<br />

c ∩ S N h = 0. Revising the argument of<br />

this inner-product form (914), we get another flavor<br />

D ( [<br />

−VN TDV )<br />

N =<br />

0<br />

δ ( −VN TDV )<br />

N<br />

] [<br />

1 T + 1 0 δ ( −VN TDV ) ] T<br />

N<br />

and we obtain mutual inversion of operators V N and D , for D ∈ S N h<br />

[ 0 0<br />

T<br />

− 2<br />

0 −VN TDV N<br />

(917)<br />

]<br />

or<br />

D = D ( V N (D) ) (918)<br />

−V T NDV N = V N<br />

(<br />

D(−V<br />

T<br />

N DV N ) ) (919)<br />

S N h = D ( V N (S N h ) ) (920)<br />

−V T N S N h V N = V N<br />

(<br />

D(−V<br />

T<br />

N S N h V N ) ) (921)<br />

Substituting Θ T Θ ← Φ into inner-product form EDM definition (863),<br />

any EDM may be expressed by the new flavor<br />

[ ]<br />

D(Φ) =<br />

∆ 0<br />

1<br />

δ(Φ)<br />

T + 1 [ 0 δ(Φ) ] [ ] 0 0 T T<br />

− 2<br />

∈ EDM N<br />

0 Φ<br />

⇔<br />

(922)<br />

Φ ≽ 0<br />

where this D is a linear surjective operator onto EDM N by definition,<br />

injective because it has no nullspace on domain S N−1<br />

+ . More broadly,<br />

aff D(S N−1<br />

+ )= D(aff S N−1<br />

+ ) (118),<br />

S N h = D(S N−1 )<br />

S N−1 = V N (S N h )<br />

(923)<br />

demonstrably isomorphisms, and by bijectivity of this inner-product form:<br />

such that<br />

EDM N = D(S N−1<br />

+ ) (924)<br />

S N−1<br />

+ = V N (EDM N ) (925)<br />

N(T(V N )) = N(V) ⊇ N(V N ) ⊇ S N⊥<br />

c<br />

= N(V)<br />

where the equality S N⊥<br />

c = N(V) is known (E.7.2.0.2).

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

Saved successfully!

Ooh no, something went wrong!