The NORM function computes the norm of a vector or a two-dimensional array.
By default, NORM computes the L2 (Euclidean) norm for vectors, and the L• norm for arrays. You may use the LNORM keyword to specify different norms.
This routine is written in the IDL language. Its source code can be found in the file norm.pro in the lib subdirectory of the IDL distribution.
Examples
A = [COMPLEX(1, 0), COMPLEX(2,-2), COMPLEX(-3,1)]
PRINT, 'Euclidian Norm of A =', NORM(A)
B = [[COMPLEX(1, 0), COMPLEX(2,-2), COMPLEX(-3,1)], $
[COMPLEX(1,-2), COMPLEX(2, 2), COMPLEX(1, 0)]]
PRINT, 'Infinity Norm of B =', NORM(B, /DOUBLE)
IDL prints:
Euclidian Norm of A = 4.35890
Infinity Norm of B = 6.9907048
Syntax
Result = NORM( A [, /DOUBLE] [, LNORM={0 | 1 | 2 | n}])
Return Value
Returns the Euclidean or infinity norm of a vector or an array. This function always returns a float or double value.
Arguments
A
A can be either a real or complex vector, or a real or complex two-dimensional array.
Keywords
DOUBLE
Set this keyword to force the result to be returned as double precision. The default is to return a single-precision result if the input is single precision, or double precision otherwise. Internally, all computations are done using double-precision arithmetic.
LNORM
Set this keyword to indicate which norm to compute. If A is a vector, then the possible values of this keyword are:
Value |
Description |
0 |
Compute the Lα norm, defined as MAX(ABS(A)).
|
1 |
Compute the L1 norm, defined as TOTAL(ABS(A)).
|
2 |
Compute the L2 norm, defined as SQRT(TOTAL(ABS(A)^2)). This is the default.
|
n
|
Compute the Ln norm, defined as (TOTAL(ABS(A)^n))^(1/n) where n is any number, float-point or integer.
|
If A is a two-dimensional array, then the possible values of this keyword are:
Value |
Description |
0 |
Compute the Lα norm (the maximum absolute row sum norm),
Defined as MAX(TOTAL(ABS(A), 1)). This is the default.
|
1 |
Compute the L1 norm (the maximum absolute column sum norm), defined as MAX(TOTAL(ABS(A), 2)).
|
2 |
Compute the L2 norm (the spectral norm) defined as the largest singular value, computed from LA_SVD.
|
Version History
See Also
COND, LA_SVD