2- norm: ║ a ║ 2 = the maximum singular value of a = (max {λ I (a h * a)}) {1/2}

(Euclid norm, the spectral norm, that is, the square root of λ 1, is the largest eigenvalue λi of A'A, where A H is the transposed * * * yoke matrix of A).

Norm is a basic concept in mathematics. In functional analysis, a norm is a function defined in a normed linear space, and any function that meets the corresponding conditions can be called a norm.

Matrix norms derived from three commonly used p- norms are:

1- norm: ║A║ 1 = max{ ∑|ai 1|, ∑|ai2|, ..., ∑|ain|} (column sum norm, the maximum sum of absolute values of elements in each column).

(where ∑|ai 1| sum of absolute values of the elements in the first column ∑| ai1| = | a1|+| A 21|+... +| an1|)

2- norm: ║ a ║ 2 = the maximum singular value of a = (max{ λi(AH*A) }) 1/2 (spectral norm, that is, the square root of λ 1 in λi of AH*A, where AH is the transposition of A * *.

∞-norm: ║A║∞ = max{ ∑|a 1j|, ∑|a2j|, ..., ∑|amj|} (row sum norm, the maximum sum of absolute values of elements in each row)