Please note that the recommended version of Scilab is 6.1.1. This page might be outdated.

However, this page did not exist in the previous stable version.

# spec

eigenvalues of matrices and pencils

### Syntax

evals=spec(A) [R,diagevals]=spec(A) evals=spec(A,B) [alpha,beta]=spec(A,B) [alpha,beta,Z]=spec(A,B) [alpha,beta,Q,Z]=spec(A,B)

### Arguments

- A
real or complex square matrix

- B
real or complex square matrix with same dimensions as

`A`

- evals
real or complex vector, the eigenvalues

- diagevals
real or complex diagonal matrix (eigenvalues along the diagonal)

- alpha
real or complex vector, al./be gives the eigenvalues

- beta
real vector, al./be gives the eigenvalues

- R
real or complex invertible square matrix, matrix right eigenvectors.

- L
real or complex invertible square matrix, pencil left eigenvectors.

- R
real or complex invertible square matrix, pencil right eigenvectors.

### Description

- evals=spec(A)
returns in vector

`evals`

the eigenvalues.- [R,diagevals] =spec(A)
returns in the diagonal matrix

`evals`

the eigenvalues and in`R`

the right eigenvectors.- evals=spec(A,B)
returns the spectrum of the matrix pencil A - s B, i.e. the roots of the polynomial matrix s B - A.

- [alpha,beta] = spec(A,B)
returns the spectrum of the matrix pencil

`A- s B`

,i.e. the roots of the polynomial matrix`A - s B`

.Generalized eigenvalues alpha and beta are so that the matrix`A - alpha./beta B`

is a singular matrix. The eigenvalues are given by`al./be`

and if`beta(i) = 0`

the ith eigenvalue is at infinity. (For`B = eye(A), alpha./beta`

is`spec(A)`

). It is usually represented as the pair (alpha,beta), as there is a reasonable interpretation for beta=0, and even for both being zero.- [alpha,beta,R] = spec(A,B)
returns in addition the matrix

`R`

of generalized right eigenvectors of the pencil.- [al,be,L,R] = spec(A,B)
returns in addition the matrix

`L`

and`R`

of generalized left and right eigenvectors of the pencil.- [al,be,Z] = spec(A,E)
returns the matrix

`Z`

of right generalized eigen vectors.- [al,be,Q,Z] = spec(A,E)
returns the matrices

`Q`

and`Z`

of right and left generalized eigen vectors.

For big full / sparse matrix, you can use the Arnoldi module.

### References

Matrix eigenvalues computations are based on the Lapack routines

DGEEV and ZGEEV when the matrix are not symmetric,

DSYEV and ZHEEV when the matrix are symmetric.

A complex symmetric matrix has conjugate offdiagonal terms and real diagonal terms.

Pencil eigenvalues computations are based on the Lapack routines DGGEV and ZGGEV.

### Real and complex matrices

It must be noticed that the type of the output variables, such as evals or R for example, is not necessarily the same as the type of the input matrices A and B. In the following paragraph, we analyse the type of the output variables in the case where one computes the eigenvalues and eigenvectors of one single matrix A.

Real A matrix

Symmetric

The eigenvalues and the eigenvectors are real.

Not symmetric

The eigenvalues and eigenvectors are complex.

Complex A matrix

Symmetric

The eigenvalues are real but the eigenvectors are complex.

Not symmetric

The eigenvalues and the eigenvectors are complex.

### Examples

// MATRIX EIGENVALUES A=diag([1,2,3]); X=rand(3,3); A=inv(X)*A*X; spec(A) x=poly(0,'x'); pol=det(x*eye(3,3)-A) roots(pol) [S,X]=bdiag(A); clean(inv(X)*A*X) // PENCIL EIGENVALUES A=rand(3,3); [al,be,R] = spec(A,eye(A)); al./be clean(inv(R)*A*R) //displaying the eigenvalues (generic matrix) A=A+%i*rand(A); E=rand(A); roots(det(A-%s*E)) //complex case

### See also

- poly — polynomial definition
- det — determinant
- schur — [ordered] Schur decomposition of matrix and pencils
- bdiag — block diagonalization, generalized eigenvectors
- colcomp — column compression, kernel, nullspace
- dsaupd — Interface for the Implicitly Restarted Arnoldi Iteration, to compute approximations to a few eigenpairs of a real and symmetric linear operator This function is obsolete. Please use eigs
- dnaupd — Interface for the Implicitly Restarted Arnoldi Iteration, to compute approximations to a few eigenpairs of a real linear operator This function is obsolete. Please use eigs

## Comments

Add a comment:Please login to comment this page.