Evanalysis
3.5Estimated reading time: 4 min

3.5 Block matrices

Partition a large matrix into smaller pieces and carry out the familiar operations block by block.

Course contents

MATH1030: Linear algebra I

Linear algebra notes.

37 sections

When matrices become large, writing every entry can hide the main idea. Block matrices fix that problem by letting you treat a large matrix as a small matrix whose "entries" are themselves smaller matrices.

This does not change the mathematics. It changes the level at which you look at the structure.

Intuition first: one large matrix, several smaller pieces

A block matrix begins with a partition. You draw horizontal and vertical cuts through a matrix and group the resulting submatrices into blocks.

The important point is that the blocks are not random decorations. They are chosen so that the shapes line up in a useful way.

Once the partition is fixed, you can often perform addition, scalar multiplication, and multiplication using whole blocks at once.

Definition

Definition

Block matrix

A block matrix is a matrix partitioned into smaller rectangular submatrices, called blocks.

You may add or subtract two block matrices block by block only when they have the same block partition.

You may multiply block matrices by the usual block formula only when the block sizes match in the same way ordinary matrix multiplication requires matching inner dimensions.

Why partitions matter

Students sometimes think that once two large matrices have the same outer size, they can automatically be added block by block. That is false.

The partition itself must match. If one matrix is cut into a 2+32 + 3 row split and the other is cut into a 1+41 + 4 row split, then the blocks do not correspond, even if both whole matrices are 5×55 × 5.

So with block matrices you always check two levels:

  1. the size of the whole matrix;
  2. the size of each block.

Worked example

Worked example

Read the 2 × 2 block multiplication rule

Suppose

A=[A11A12A21A22],B=[B11B12B21B22].A = \begin{bmatrix} A_{11} & A_{12} \\ A_{21} & A_{22} \end{bmatrix}, \qquad B = \begin{bmatrix} B_{11} & B_{12} \\ B_{21} & B_{22} \end{bmatrix}.

If the block sizes are compatible, then

AB=[A11B11+A12B21A11B12+A12B22A21B11+A22B21A21B12+A22B22].AB = \begin{bmatrix} A_{11}B_{11} + A_{12}B_{21} & A_{11}B_{12} + A_{12}B_{22} \\ A_{21}B_{11} + A_{22}B_{21} & A_{21}B_{12} + A_{22}B_{22} \end{bmatrix}.

This looks exactly like ordinary 2×22 × 2 matrix multiplication, except each "entry" is now a whole submatrix.

Why block multiplication is useful

Block multiplication is not a new algebraic universe. It is a bookkeeping tool for ordinary matrix multiplication.

Its value is that it lets you:

  • separate variables into groups,
  • keep related rows and columns together,
  • reuse familiar small patterns inside a large computation.

This becomes especially useful later when matrices naturally split into coefficient blocks, identity blocks, or transformation blocks.

Common mistakes

Common mistake

Assuming any partition will do

Block formulas only work when the partitions are compatible. The cuts must be chosen so that the relevant block products are defined.

Common mistake

Forgetting that block operations still obey ordinary size rules

Each block product such as A12B21A_{12}B_{21} still needs matching inner dimensions. The block view does not cancel the usual multiplication rule.

Quick checks

Quick check

Why do two block matrices need the same partition before you add them block by block?

Answer using the phrase "corresponding blocks."

Solution

Answer

Quick check

What is the main idea behind the formula for 2 × 2 block multiplication?

Do not compute; describe the pattern.

Solution

Answer

Exercises

Quick check

Explain why block multiplication is still ordinary matrix multiplication in disguise.

Use the words "same rule" somewhere in your answer.

Solution

Guided solution

Quick check

A pair of 4×44 × 4 matrices have different block cuts. Can you still add them block by block? Why or why not?

Answer in one or two sentences.

Solution

Guided solution

Review 3.2 Matrix multiplication and linear systems if the ordinary multiplication pattern is still shaky.

Continue to 5.1 Invertible matrices when you are ready to see how matrix structure connects to reversibility.

Premium learning add-ons

Core notes stay free. Advanced exercises, video explanations, and premium exports are available through paid plans.