10.03.2015 Views

v2009.01.01 - Convex Optimization

v2009.01.01 - Convex Optimization

v2009.01.01 - Convex Optimization

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

668 APPENDIX E. PROJECTION<br />

By (1850), product P 1 XP 1 is the one-dimensional orthogonal projection of<br />

X in isomorphic R m2 on the range of vectorized P 1 because, for rankP 1 =1<br />

and P 2<br />

1 =P 1 ∈ S m (confer (1842))<br />

P 1 XP 1 = yT Xy<br />

y T y<br />

〈 〉<br />

yy T yy<br />

T yy<br />

T<br />

y T y = y T y , X y T y = 〈P 1 , X〉 P 1 = 〈P 1 , X〉<br />

〈P 1 , P 1 〉 P 1<br />

(1864)<br />

The coefficient of orthogonal projection 〈P 1 , X〉= y T Xy/(y T y) is also known<br />

as Rayleigh’s quotient. E.12 When P 1 is rank-one symmetric as in (1863),<br />

R(vecP 1 XP 1 ) = R(vec P 1 ) in R m2 (1865)<br />

and<br />

P 1 XP 1 − X ⊥ P 1 in R m2 (1866)<br />

The test for positive semidefiniteness, then, is a test for nonnegativity of<br />

the coefficient of orthogonal projection of X on the range of each and every<br />

vectorized extreme direction yy T (2.8.1) from the positive semidefinite cone<br />

in the ambient space of symmetric matrices.<br />

E.12 When y becomes the j th eigenvector s j of diagonalizable X , for example, 〈P 1 , X 〉<br />

becomes the j th eigenvalue: [170,III]<br />

〈P 1 , X 〉| y=sj<br />

=<br />

s T j<br />

( m∑<br />

λ i s i wi<br />

T<br />

i=1<br />

s T j s j<br />

)<br />

s j<br />

= λ j<br />

Similarly for y = w j , the j th left-eigenvector,<br />

〈P 1 , X 〉| y=wj<br />

=<br />

w T j<br />

( m∑<br />

λ i s i wi<br />

T<br />

i=1<br />

w T j w j<br />

)<br />

w j<br />

= λ j<br />

A quandary may arise regarding the potential annihilation of the antisymmetric part of<br />

X when s T j Xs j is formed. Were annihilation to occur, it would imply the eigenvalue thus<br />

found came instead from the symmetric part of X . The quandary is resolved recognizing<br />

that diagonalization of real X admits complex eigenvectors; hence, annihilation could only<br />

come about by forming Re(s H j Xs j) = s H j (X +XT )s j /2 [176,7.1] where (X +X T )/2 is<br />

the symmetric part of X , and s H j denotes conjugate transpose.

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

Saved successfully!

Ooh no, something went wrong!