cdf2rdf

Convert complex diagonal form to real block diagonal form

Syntax

[Vnew,Dnew] = cdf2rdf(V,D)

Description

[Vnew,Dnew] = cdf2rdf(V,D) transforms the outputs of [V,D] = eig(X) or [V,D] = eigs(X,___) for real matrices X from complex diagonal form to real diagonal form. This operation transforms how the eigenvalues of X are expressed in D, and transforms V such that X*Vnew = Vnew*Dnew.

• In complex diagonal form, D is a diagonal matrix with complex conjugate pairs of eigenvalues on the main diagonal:

  $$\left[\begin{array}{cccccc}{\lambda }_1& & & & & \\ & a+bi& & & & \\ & & a-bi& & & \\ & & & c+di& & \\ & & & & c-di& \\ & & & & & \ddots \end{array}\right]$$

  Some of the eigenvalues along the diagonal might be real, but complex conjugate eigenvalue pairs are assumed to be next to one another.

• In real diagonal form, Dnew has real eigenvalues on the diagonal, and complex eigenvalues are expressed as 2-by-2 real blocks along the main diagonal:

  $$\left[\begin{array}{cccccc}{\lambda }_1& & & & & \\ & a& b& & & \\ & -b& a& & & \\ & & & c& d& \\ & & & -d& c& \\ & & & & & \ddots \end{array}\right]$$

Examples

Transform Complex Diagonal Matrix into Real Block Diagonal Matrix

Calculate the eigenvalues and eigenvectors of a real matrix, and then transform the complex conjugate eigenvalue pairs into real blocks.

Find the eigenvalues and eigenvectors of a real matrix.

X = [1   1   1   1   1 
     0   4   5   1   1 
     0  -5   4   1   1 
     0   0   2   3   1
     0   0  -3  -2   1];
[V,D] = eig(X)

V = 5×5 complex

   1.0000 + 0.0000i   0.0179 + 0.1351i   0.0179 - 0.1351i   0.1593 - 0.4031i   0.1593 + 0.4031i
   0.0000 + 0.0000i  -0.0130 + 0.6214i  -0.0130 - 0.6214i   0.0704 - 0.0267i   0.0704 + 0.0267i
   0.0000 + 0.0000i  -0.6363 + 0.0000i  -0.6363 + 0.0000i  -0.1261 + 0.1032i  -0.1261 - 0.1032i
   0.0000 + 0.0000i  -0.1045 + 0.2087i  -0.1045 - 0.2087i  -0.2279 - 0.4161i  -0.2279 + 0.4161i
   0.0000 + 0.0000i   0.1156 - 0.3497i   0.1156 + 0.3497i   0.7449 + 0.0000i   0.7449 + 0.0000i

D = 5×5 complex

   1.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i
   0.0000 + 0.0000i   3.8801 + 5.1046i   0.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i
   0.0000 + 0.0000i   0.0000 + 0.0000i   3.8801 - 5.1046i   0.0000 + 0.0000i   0.0000 + 0.0000i
   0.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i   2.1199 + 0.7018i   0.0000 + 0.0000i
   0.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i   0.0000 + 0.0000i   2.1199 - 0.7018i

D contains one real eigenvalue and two pairs of complex conjugate eigenvalues.

Transform V and D so that Dnew is in real block diagonal form and Vnew satisfies X*Vnew = Vnew*Dnew

[Vnew,Dnew] = cdf2rdf(V,D)

Vnew = 5×5

    1.0000    0.0253    0.1911    0.2253   -0.5701
         0   -0.0184    0.8789    0.0996   -0.0378
         0   -0.8999         0   -0.1784    0.1459
         0   -0.1478    0.2951   -0.3222   -0.5885
         0    0.1634   -0.4946    1.0534         0

Dnew = 5×5

    1.0000         0         0         0         0
         0    3.8801    5.1046         0         0
         0   -5.1046    3.8801         0         0
         0         0         0    2.1199    0.7018
         0         0         0   -0.7018    2.1199

Dnew still has the real eigenvalue, but the complex conjugate eigenvalues are replaced with 2-by-2 blocks.

Input Arguments

V — Right eigenvector matrix

matrix

Right eigenvector matrix, specified as the matrix returned by [V,D] = eig(X) or [V,D] = eigs(X,___).

Data Types: single | double
Complex Number Support: Yes

D — Diagonal eigenvalue matrix

matrix

Diagonal eigenvalue matrix, specified as the matrix returned by [V,D] = eig(X) or [V,D] = eigs(X,___). Some of the eigenvalues along the diagonal of D might be real, but complex conjugate eigenvalue pairs are assumed to be next to one another.

Since eigs returns a subset of the eigenvalues and eigenvectors, the requested number of eigenvalues might include half of a complex conjugate pair. cdf2rdf returns an error if the D input contains incomplete complex conjugate pairs.

Data Types: single | double
Complex Number Support: Yes

Output Arguments

Vnew — Transformed right eigenvector matrix

matrix

Transformed right eigenvector matrix, returned as a matrix. The eigenvectors in V are transformed so that X*Vnew = Vnew*Dnew holds. If the input eigenvector matrix V is unitary, then Vnew is as well. After the transformation, the individual columns of Vnew are no longer eigenvectors of X, but each pair of vectors in Vnew associated with a 2-by-2 block in Dnew spans the corresponding invariant vectors.

Dnew — Transformed diagonal eigenvalue matrix

matrix

Transformed diagonal eigenvalue matrix, returned as a block diagonal real matrix. Complex conjugate eigenvalue pairs in D are replaced with 2-by-2 real blocks along the diagonal in Dnew.

Extended Capabilities

Thread-Based Environment

Run code in the background using MATLAB® backgroundPool or accelerate code with Parallel Computing Toolbox™ ThreadPool.

This function fully supports thread-based environments. For more information, see Run MATLAB Functions in Thread-Based Environment.

GPU Arrays

Accelerate code by running on a graphics processing unit (GPU) using Parallel Computing Toolbox™.

The cdf2rdf function fully supports GPU arrays. To run the function on a GPU, specify the input data as a gpuArray (Parallel Computing Toolbox). For more information, see Run MATLAB Functions on a GPU (Parallel Computing Toolbox).

Version History

Introduced before R2006a

R2021a: Improved algorithm to reduce round-off error

The algorithm for all input matrices has been improved to reduce floating-point round-off errors in the calculation.