Multivariable-Calculus ยท Unit 3 ยท Video 3 ยท Interactive Practice

Rows Dot Columns: Packaging a Change of Coordinates as AX = U

IKey Formulas

FormulaNameWhat you need
ui=ai1x1+ai2x2+ai3x3u_i = a_{i1}x_1 + a_{i2}x_2 + a_{i3}x_3One line of the systemThe coefficients of the iith equation
AX=UAX = UMatrix formCoefficient table AA, input column XX
(AB)ij=ai1b1j+ai2b2j+โ‹ฏ+ainbnj(AB)_{ij} = a_{i1}b_{1j} + a_{i2}b_{2j} + \cdots + a_{in}b_{nj}Entry of a productRow ii of AA, column jj of BB
(mร—n)(nร—p)=mร—p(m \times n)(n \times p) = m \times pShape ruleWidth of AA must equal height of BB

Key Insight: One row, one column, one dot product, one entry โ€” every other rule about matrix products follows from that one.

IIVisualization 1 โ€” One Row Makes One Output

Each line of the system is one row of AA dotted with the column XX.

IIIVisualization 2 โ€” Which Row, Which Column?

An entry's position names its ingredients: the row to its left, the column above it.

๐Ÿ’ก Order is not negotiable: BABA would pair BB's rows with AA's columns, and with these shapes it does not even exist.

IVVisualization 3 โ€” When Does ABAB Exist?

A dot product needs two lists of the same length, and that single fact decides everything.

VQuiz Questions

Problem 1 ยท One Entry of a Product

Given: the first row of AA is (2โˆ’13)\begin{pmatrix} 2 & -1 & 3 \end{pmatrix} and the second column of BB is (465)\begin{pmatrix} 4 \\ 6 \\ 5 \end{pmatrix} โ€” find the entry (AB)12(AB)_{12}.

โœ… Correct! 2(4)+(โˆ’1)(6)+3(5)=8โˆ’6+15=172(4) + (-1)(6) + 3(5) = 8 - 6 + 15 = 17, and it sits in row 1, column 2 because it used row 1 of AA and column 2 of BB.
โŒ Close, but check the sign. The second entry of the row is โˆ’1-1, so that term is โˆ’6-6, not +6+6.
โŒ Not quite. The two lists pair up in order โ€” first with first, second with second, third with third โ€” not front to back.
โŒ Not quite. A dot product multiplies the matched entries first and adds the products; it is not the sum of the six numbers.
Show solution

The entry in row 11, column 22 of ABAB is row 11 of AA dotted with column 22 of BB:

(AB)12=2(4)+(โˆ’1)(6)+3(5)=8โˆ’6+15=17(AB)_{12} = 2(4) + (-1)(6) + 3(5) = 8 - 6 + 15 = 17

Both lists have three entries, so every term of the row has a partner in the column โ€” which is exactly the condition that lets the product exist.

Problem 2 ยท Packing a System into AX=UAX = U

Given: the system u1=4x2+2x1u_1 = 4x_2 + 2x_1, ย u2=x3โˆ’3x1\ u_2 = x_3 - 3x_1, ย u3=x1+5x3\ u_3 = x_1 + 5x_3 โ€” find the matrix AA for which AX=UAX = U.

โœ… Correct! Row ii holds the coefficients of x1,x2,x3x_1, x_2, x_3 in equation ii, in that order, with 00 wherever a variable is absent.
โŒ Not quite. A column of AA belongs to one variable, so the coefficients go in the order x1,x2,x3x_1, x_2, x_3 โ€” not in the order the terms happen to be written.
โŒ Not quite. That is the table with rows and columns swapped: row ii must reproduce equation ii when dotted with XX, so equation 2 belongs in row 2, not column 2.
โŒ Close, but check a sign. In u2=x3โˆ’3x1u_2 = x_3 - 3x_1 the coefficient of x1x_1 is โˆ’3-3.
Show solution

Rewrite each equation in the order x1,x2,x3x_1, x_2, x_3 with every coefficient written out โ€” including the 00s and 11s that are usually left implicit:

u1=2x1+4x2+0x3,u2=โˆ’3x1+0x2+1x3,u3=1x1+0x2+5x3u_1 = 2x_1 + 4x_2 + 0x_3, \qquad u_2 = -3x_1 + 0x_2 + 1x_3, \qquad u_3 = 1x_1 + 0x_2 + 5x_3

Lifting the coefficients out, keeping their positions:

A=(240โˆ’301105)A = \begin{pmatrix} 2 & 4 & 0 \\ -3 & 0 & 1 \\ 1 & 0 & 5 \end{pmatrix}

Check row 2: (โˆ’3)x1+0x2+1x3=x3โˆ’3x1=u2(-3)x_1 + 0x_2 + 1x_3 = x_3 - 3x_1 = u_2 โœ“

Problem 3 ยท Shapes First

Given: AA is a 2ร—52 \times 5 matrix and BB is a 5ร—35 \times 3 matrix โ€” find the size of each product, or say it does not exist.

Size of ABAB

Size of BABA

โœ… Correct! (2ร—5)(5ร—3)(2 \times 5)(5 \times 3) has matching inner numbers and gives 2ร—32 \times 3; (5ร—3)(2ร—5)(5 \times 3)(2 \times 5) does not, so BABA is undefined.
โŒ The two sizes are the wrong way round. The answer inherits the height of the left matrix and the width of the right one.
โŒ Check the inner numbers. A row of AA has 55 entries and a column of BB has 55 entries, so the dot products do exist.
โŒ Check the inner numbers. In BABA a row of BB has 33 entries while a column of AA has only 22.
Show solution

ABAB: line up the shapes as (2ร—5)(5ร—3)(2 \times 5)(5 \times 3). The inner numbers are 55 and 55, so every row of AA and every column of BB are lists of the same length and the product exists. Its size is the outer pair:

(2ร—5)(5ร—3)=2ร—3(2 \times 5)(5 \times 3) = 2 \times 3

BABA: now the shapes line up as (5ร—3)(2ร—5)(5 \times 3)(2 \times 5). A row of BB has 33 entries, a column of AA has 22; the third entry of the row has nothing to pair with, so BABA does not exist.

Reversing the factors changes the question entirely โ€” matrix multiplication is not commutative, and the two products need not even have the same shape.

Problem 4 ยท The Change of Coordinates (Video Example)

Given: the video's matrix A=(233245112)A = \begin{pmatrix} 2 & 3 & 3 \\ 2 & 4 & 5 \\ 1 & 1 & 2 \end{pmatrix} and the point with old coordinates X=(21โˆ’1)X = \begin{pmatrix} 2 \\ 1 \\ -1 \end{pmatrix} โ€” find the new coordinates U=AXU = AX.

โœ… Correct! Three rows, three dot products, three new coordinates โ€” all three equations at once.
โŒ Close, but check the sign. The third old coordinate is โˆ’1-1, so each third term subtracts: 3(โˆ’1)=โˆ’33(-1) = -3, 5(โˆ’1)=โˆ’55(-1) = -5, 2(โˆ’1)=โˆ’22(-1) = -2.
โŒ Not quite. Those are the columns of AA dotted with XX. Each entry of AXAX uses a row of AA โ€” the coefficients of one equation.
โŒ Not quite. Each output uses the entire row and the entire column: all three terms, not just the x1x_1 term.
Show solution

Dot each row of AA with the column X=(2,โ€‰1,โ€‰โˆ’1)X = (2,\, 1,\, -1):

u1=2(2)+3(1)+3(โˆ’1)=4+3โˆ’3=4u_1 = 2(2) + 3(1) + 3(-1) = 4 + 3 - 3 = 4 u2=2(2)+4(1)+5(โˆ’1)=4+4โˆ’5=3u_2 = 2(2) + 4(1) + 5(-1) = 4 + 4 - 5 = 3 u3=1(2)+1(1)+2(โˆ’1)=2+1โˆ’2=1u_3 = 1(2) + 1(1) + 2(-1) = 2 + 1 - 2 = 1

So U=(431)U = \begin{pmatrix} 4 \\ 3 \\ 1 \end{pmatrix}: the same point, now measured in the new axes.

The shapes agree with the rule: (3ร—3)(3ร—1)=3ร—1(3 \times 3)(3 \times 1) = 3 \times 1, a column of three new coordinates.

Solved: 0 / 4