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

648 APPENDIX B. SIMPLE MATRICESEach and every eigenvalue of a (real) orthogonal matrix has magnitude 1λ(Q) ∈ C n , |λ(Q)| = 1 (1664)while only the identity matrix can be simultaneously positive definite andorthogonal.B.5.2ReflectionA matrix for pointwise reflection is defined by imposing symmetry uponthe orthogonal matrix; id est, a reflection matrix is completely defined byQ −1 = Q T = Q . The reflection matrix is a symmetric orthogonal matrix,and vice versa, characterized:Q T = Q , Q −1 = Q T , ‖Q‖ 2 = 1 (1665)The Householder matrix (B.3.1) is an example of symmetric orthogonal(reflection) matrix.Reflection matrices have eigenvalues equal to ±1 and so detQ=±1. Itis natural to expect a relationship between reflection and projection matricesbecause all projection matrices have eigenvalues belonging to {0, 1}. Infact, any reflection matrix Q is related to some orthogonal projector P by[204,1 prob.44]Q = I − 2P (1666)Yet P is, generally, neither orthogonal or invertible. (E.3.2)λ(Q) ∈ R n , |λ(Q)| = 1 (1667)Reflection is with respect to R(P ) ⊥ . Matrix 2P −I represents antireflection.Every orthogonal matrix can be expressed as the product of a rotation anda reflection. The collection of all orthogonal matrices of particular dimensiondoes not form a convex set.B.5.3Rotation of range and rowspaceGiven orthogonal matrix Q , column vectors of a matrix X are simultaneouslyrotated about the origin via product QX . In three dimensions (X ∈ R 3×N ),

B.5. ORTHOGONAL MATRIX 649Figure 159: Gimbal: a mechanism imparting three degrees of dimensionalfreedom to a Euclidean body suspended at the device’s center. Each ring isfree to rotate about one axis. (Courtesy of The MathWorks Inc.)the precise meaning of rotation is best illustrated in Figure 159 where thegimbal aids visualization of what is achievable; mathematically, (5.5.2.0.1)⎡⎤⎡⎤⎡⎤cosθ 0 − sin θ 1 0 0 cos φ − sin φ 0Q = ⎣ 0 1 0 ⎦⎣0 cos ψ − sin ψ ⎦⎣sin φ cos φ 0 ⎦ (1668)sin θ 0 cos θ 0 sinψ cos ψ 0 0 1B.5.3.0.1 Example. One axis of revolution. [ RPartition an n+1-dimensional Euclidean space R n+1 nRan n-dimensional subspace]and defineR {λ∈ R n+1 | 1 T λ = 0} (1669)(a hyperplane through the origin). We want an orthogonal matrix thatrotates a list in the columns of matrix X ∈ R n+1×N through the dihedralangle between R n and R (2.4.3)( ) ( )(R n 〈en+1 , 1〉1, R) = arccos = arccos √ radians (1670)‖e n+1 ‖ ‖1‖ n+1The vertex-description of the nonnegative orthant in R n+1 is{[e 1 e 2 · · · e n+1 ]a | a ≽ 0} = {a ≽ 0} = R n+1+ ⊂ R n+1 (1671)

B.5. ORTHOGONAL MATRIX 649Figure 159: Gimbal: a mechanism imparting three degrees of dimensionalfreedom to a Euclidean body suspended at the device’s center. Each ring isfree to rotate about one axis. (Courtesy of The MathWorks Inc.)the precise meaning of rotation is best illustrated in Figure 159 where thegimbal aids visualization of what is achievable; mathematically, (5.5.2.0.1)⎡⎤⎡⎤⎡⎤cosθ 0 − sin θ 1 0 0 cos φ − sin φ 0Q = ⎣ 0 1 0 ⎦⎣0 cos ψ − sin ψ ⎦⎣sin φ cos φ 0 ⎦ (1668)sin θ 0 cos θ 0 sinψ cos ψ 0 0 1B.5.3.0.1 Example. One axis of revolution. [ RPartition an n+1-dimensional Euclidean space R n+1 nRan n-dimensional subspace]and defineR {λ∈ R n+1 | 1 T λ = 0} (1669)(a hyperplane through the origin). We want an orthogonal matrix thatrotates a list in the columns of matrix X ∈ R n+1×N through the dihedralangle between R n and R (2.4.3)( ) ( )(R n 〈en+1 , 1〉1, R) = arccos = arccos √ radians (1670)‖e n+1 ‖ ‖1‖ n+1The vertex-description of the nonnegative orthant in R n+1 is{[e 1 e 2 · · · e n+1 ]a | a ≽ 0} = {a ≽ 0} = R n+1+ ⊂ R n+1 (1671)

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

Saved successfully!

Ooh no, something went wrong!