watch: LA - 高山 08 | Properties of matrix multiplication

Matrix multiplication doesn’t have commutative property (亀捒律). But some square matrices satisfy 𝐀𝐁=𝐁𝐀

Matrix multiplication has Associative property (η»“εˆεΎ‹)

(𝐀𝐁)𝐂 = 𝐀(𝐁𝐂)

Compund linear mapping

Transpose and multiplication

(𝐀𝐁)α΅€ β‰  𝐀ᡀ𝐁ᡀ

(𝐀𝐁)α΅€ = 𝐁ᡀ𝐀ᡀ

Square matrix multiplication

The number of rows = the number of columns

𝐀 = $[^{1\ 2} _{1\ 4}]$ ; 𝐁 = $[^{4\ -2} _{-1\ 1}]$

There is 𝐀𝐁 = 𝐁𝐀.

So these two matrices 𝐀, 𝐁 are commutative 1. 𝐀,𝐁 must be the square matrices with the same size, and the result also has the identical-size square matrix.

Power of square matrix

𝐀ᡏ = 𝐀⋅𝐀ᡏ⁻¹ = 𝐀ᡏ⁻¹⋅𝐀, where 𝐀 and 𝐀ᡏ⁻¹ are commutative because of the commutative property of matrix multiplication.

Therefore, the two positive integar powers of matrix 𝐀 are always commutative.

Power of the matrix only applicable to square matrices. Otherwise, the mismatched dimensions prevent performing matrices multiplication.

𝐀¹=𝐀 ; 𝐀⁰ = 𝐈

Thus, the range of power k is the set of natural numbers k∈ β„•

Power of diagonal matrix

$$ 𝐀 = \begin{bmatrix} 1 & 0 & 0 \\ 0 & 2 & 0 \\ 0 & 0 & 3 \end{bmatrix} = diag\{1, 2, 3\} $$

Each elements on the main diagonal is multiplied with itself many times.

2

Scalar matrix

$$ 𝐀 = \begin{bmatrix} Ξ» & 0 & 0 \\ 0 & Ξ» & 0 \\ 0 & 0 & Ξ» \end{bmatrix} = diag\{Ξ», Ξ», Ξ»\} $$

If Ξ» = 1, then 𝐀 = 𝐈. So 𝐈ᡏ=𝐈, and this is also applicable to zero matrix 𝟎ᡏ= 𝟎 (kβ‰ 0)


Ref