v2010.10.26 - Convex Optimization

v2010.10.26 - Convex Optimization v2010.10.26 - Convex Optimization

convexoptimization.com
from convexoptimization.com More from this publisher
12.07.2015 Views

728 APPENDIX E. PROJECTION(A.2) It is a fact that y T Xy is always proportional to a coefficient oforthogonal projection; letting z in formula (1975) become y ∈ R m , thenP 2 =P 1 =yy T /y T y=yy T /‖yy T ‖ 2 (confer (1610)) and formula (1976) becomes〈yy T , X〉〈yy T , yy T 〉 yyT = yT Xyy T yyy Ty T y = yyTy T y X yyTy T y P 1XP 1 (1987)By (1974), product P 1 XP 1 is the one-dimensional orthogonal projection ofX in isomorphic R m2 on the range of vectorized P 1 because, for rankP 1 =1and P 21 =P 1 ∈ S m (confer (1966))P 1 XP 1 = yT Xyy T y〈 〉yy T yyT yyTy T y = y T y , X y T y = 〈P 1 , X〉 P 1 = 〈P 1 , X〉〈P 1 , P 1 〉 P 1(1988)The coefficient of orthogonal projection 〈P 1 , X〉= y T Xy/(y T y) is also knownas Rayleigh’s quotient. E.12 When P 1 is rank-one symmetric as in (1987),R(vec P 1 XP 1 ) = R(vec P 1 ) in R m2 (1989)andP 1 XP 1 − X ⊥ P 1 in R m2 (1990)E.12 When y becomes the j th eigenvector s j of diagonalizable X , for example, 〈P 1 , X 〉becomes the j th eigenvalue: [195,III]( m∑)s T j λ i s i wiT s ji=1〈P 1 , X 〉| y=sj=s T j s = λ jjSimilarly for y = w j , the j th left-eigenvector,( m∑〈P 1 , X 〉| y=wj=w T jλ i s i wiTi=1wj Tw j)w j= λ jA quandary may arise regarding the potential annihilation of the antisymmetric part ofX when s T j Xs j is formed. Were annihilation to occur, it would imply the eigenvalue thusfound came instead from the symmetric part of X . The quandary is resolved recognizingthat diagonalization of real X admits complex eigenvectors; hence, annihilation could onlycome about by forming re(s H j Xs j) = s H j (X +XT )s j /2 [202,7.1] where (X +X T )/2 isthe symmetric part of X , and s H j denotes conjugate transpose.

E.7. PROJECTION ON MATRIX SUBSPACES 729The test for positive semidefiniteness, then, is a test for nonnegativity ofthe coefficient of orthogonal projection of X on the range of each and everyvectorized extreme direction yy T (2.8.1) from the positive semidefinite conein the ambient space of symmetric matrices.E.6.4.3 PXP ≽ 0In some circumstances, it may be desirable to limit the domain of testy T Xy ≥ 0 for positive semidefiniteness; e.g., {‖y‖= 1}. Another exampleof limiting domain-of-test is central to Euclidean distance geometry: ForR(V )= N(1 T ) , the test −V DV ≽ 0 determines whether D ∈ S N h is aEuclidean distance matrix. The same test may be stated: For D ∈ S N h (andoptionally ‖y‖=1)D ∈ EDM N ⇔ −y T Dy = 〈yy T , −D〉 ≥ 0 ∀y ∈ R(V ) (1991)The test −V DV ≽ 0 is therefore equivalent to a test for nonnegativity of thecoefficient of orthogonal projection of −D on the range of each and everyvectorized extreme direction yy T from the positive semidefinite cone S N + suchthat R(yy T )= R(y)⊆ R(V ). (The validity of this result is independent ofwhether V is itself a projection matrix.)E.7 Projection on matrix subspacesE.7.1 PXP misinterpretation for higher-rank PFor a projection matrix P of rank greater than 1, PXP is generally notcommensurate with 〈P,X 〉 P as is the case for projector dyads (1988). Yet〈P,P 〉for a symmetric idempotent matrix P of any rank we are tempted to say“ PXP is the orthogonal projection of X ∈ S m on R(vec P) ”. The fallacyis: vec PXP does not necessarily belong to the range of vectorized P ; themost basic requirement for projection on R(vec P).E.7.2Orthogonal projection on matrix subspacesWith A 1 ∈ R m×n , B 1 ∈ R n×k , Z 1 ∈ R m×k , A 2 ∈ R p×n , B 2 ∈ R n×k , Z 2 ∈ R p×k asdefined for nonorthogonal projector (1886), and definingP 1 A 1 A † 1 ∈ S m , P 2 A 2 A † 2 ∈ S p (1992)

E.7. PROJECTION ON MATRIX SUBSPACES 729The test for positive semidefiniteness, then, is a test for nonnegativity ofthe coefficient of orthogonal projection of X on the range of each and everyvectorized extreme direction yy T (2.8.1) from the positive semidefinite conein the ambient space of symmetric matrices.E.6.4.3 PXP ≽ 0In some circumstances, it may be desirable to limit the domain of testy T Xy ≥ 0 for positive semidefiniteness; e.g., {‖y‖= 1}. Another exampleof limiting domain-of-test is central to Euclidean distance geometry: ForR(V )= N(1 T ) , the test −V DV ≽ 0 determines whether D ∈ S N h is aEuclidean distance matrix. The same test may be stated: For D ∈ S N h (andoptionally ‖y‖=1)D ∈ EDM N ⇔ −y T Dy = 〈yy T , −D〉 ≥ 0 ∀y ∈ R(V ) (1991)The test −V DV ≽ 0 is therefore equivalent to a test for nonnegativity of thecoefficient of orthogonal projection of −D on the range of each and everyvectorized extreme direction yy T from the positive semidefinite cone S N + suchthat R(yy T )= R(y)⊆ R(V ). (The validity of this result is independent ofwhether V is itself a projection matrix.)E.7 Projection on matrix subspacesE.7.1 PXP misinterpretation for higher-rank PFor a projection matrix P of rank greater than 1, PXP is generally notcommensurate with 〈P,X 〉 P as is the case for projector dyads (1988). Yet〈P,P 〉for a symmetric idempotent matrix P of any rank we are tempted to say“ PXP is the orthogonal projection of X ∈ S m on R(vec P) ”. The fallacyis: vec PXP does not necessarily belong to the range of vectorized P ; themost basic requirement for projection on R(vec P).E.7.2Orthogonal projection on matrix subspacesWith A 1 ∈ R m×n , B 1 ∈ R n×k , Z 1 ∈ R m×k , A 2 ∈ R p×n , B 2 ∈ R n×k , Z 2 ∈ R p×k asdefined for nonorthogonal projector (1886), and definingP 1 A 1 A † 1 ∈ S m , P 2 A 2 A † 2 ∈ S p (1992)

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

Saved successfully!

Ooh no, something went wrong!