Linear AlgebraVector Spaces and Subspaces8 of 9
Vector Spaces and Subspaces · 8

Change of Basis

We describe vectors with coordinates on a grid, and we need a way to translate those coordinates when we switch to a different grid.

1

Translating Between a Custom Basis and the Standard Grid

To translate a vector's coordinates between our standard (x,y)(x, y) grid and a custom Basis BB built from b1=[21]\bone{b_1} = \bone{\begin{bmatrix} 2 \\ 1 \end{bmatrix}} and b2=[−12]\btwo{b_2} = \btwo{\begin{bmatrix} -1 \\ 2 \end{bmatrix}}, we put the new basis vectors into the columns of a matrix BB.

Translating Between a Custom Basis and the Standard Grid
[??]\htmlClass{lc-blue}{\begin{bmatrix} \htmlClass{lc-ink}{?} \\ \htmlClass{lc-ink}{?} \end{bmatrix}}
Text solution
Multiply by BB to convert from Basis BB to the standard gridPut the basis vectors b1\bone{b_1} and b2\btwo{b_2} into the columns of BB and multiply by [31]B\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{1} \end{bmatrix}}_B to find where 3\bone{3} steps of b1\bone{b_1} and 1\btwo{1} step of b2\btwo{b_2} land in our standard coordinates.
B[31]B=[2−112][31]B=3[21]+1[−12]=[55]ours\displaystyle \begin{aligned} B\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{1} \end{bmatrix}}_B &= \begin{bmatrix} \bone{2} & \btwo{-1} \\ \bone{1} & \btwo{2} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{1} \end{bmatrix}}_B \\ &= \bone{3}\bone{\begin{bmatrix} 2 \\ 1 \end{bmatrix}} + \btwo{1}\btwo{\begin{bmatrix} -1 \\ 2 \end{bmatrix}} \\ &= \arrowv{\begin{bmatrix} 5 \\ 5 \end{bmatrix}}_{\text{ours}} \end{aligned}
Row-reduce or multiply by B−1B^{-1} to convert from the standard grid to Basis BBTo find the BB-coordinates of the standard vector [34]ours\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}}, solve B[xy]B=[34]oursB\arrowv{\begin{bmatrix} \bone{x} \\ \btwo{y} \end{bmatrix}}_B = \arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} by row-reducing the augmented matrix or multiplying by B−1B^{-1}.
[2−13124]→row-reduce[102011]  ⟹  [xy]B=B−1[34]ours=[21]B\displaystyle \begin{aligned} &\left[\begin{array}{cc|c} \bone{2} & \btwo{-1} & \arrowv{3} \\ \bone{1} & \btwo{2} & \arrowv{4} \end{array}\right] \\ &\xrightarrow{\text{row-reduce}} \left[\begin{array}{cc|c} \bone{1} & \btwo{0} & \bone{2} \\ \bone{0} & \btwo{1} & \btwo{1} \end{array}\right] \\ &\implies \arrowv{\begin{bmatrix} \bone{x} \\ \btwo{y} \end{bmatrix}}_B = B^{-1}\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} = \arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B \end{aligned}

So [34]ours\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} and [21]B\arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B describe the exact same arrow on the board, just counted on two different grids.

2

Connecting Two Custom Bases Through the Standard Grid

Multiplying by BB or B−1B^{-1} connects a custom basis to the standard grid, but now suppose we have a second custom grid, Basis CC with c1=[11]\cbasis{c_1} = \cbasis{\begin{bmatrix} 1 \\ 1 \end{bmatrix}} and c2=[12]\cbasis{c_2} = \cbasis{\begin{bmatrix} 1 \\ 2 \end{bmatrix}}, and we want a single matrix that translates a vector's coordinates directly from Basis BB to Basis CC.

Connecting Two Custom Bases Through the Standard Grid
b1\htmlClass{lc-green}{b_1}
b2\htmlClass{lc-pink}{b_2}
c1\htmlClass{lc-violet}{c_1}
c2\htmlClass{lc-violet}{c_2}
v\htmlClass{lc-blue}{v}
[v]B=[32]B[v]C=  ?\begin{aligned} [\htmlClass{lc-blue}{v}]_B &= \htmlClass{lc-blue}{\begin{bmatrix} \htmlClass{lc-green}{3} \\ \htmlClass{lc-pink}{2} \end{bmatrix}}_B \\[0.3em] [\htmlClass{lc-blue}{v}]_C &= \;? \end{aligned}

Instead of inventing a new rule, we go through the standard grid in the middle: first multiply [v]B[\arrowv{v}]_B by BB to get standard coordinates, and then multiply that result by C−1C^{-1} to get the coordinates in Basis CC.

[v]C=C−1B⏟PC←Bchange-of-basis matrix [v]B\displaystyle \begin{aligned} [\arrowv{v}]_C = \underbrace{C^{-1}B}_{\substack{\\[-8pt] P_{C \leftarrow B} \\[3pt] \text{change-of-basis matrix}}}\,[\arrowv{v}]_B \end{aligned}

The subscript arrow in PC←BP_{C \leftarrow B} points from right to left to match the order of matrix multiplication: the input vector [v]B[\arrowv{v}]_B goes in on the right, and the translated output [v]C[\arrowv{v}]_C comes out on the left.

Basis B\text{Basis }B
standard grid\text{standard grid}
Basis C\text{Basis }C
[32]B\htmlClass{lc-blue}{\begin{bmatrix} \htmlClass{lc-green}{3} \\ \htmlClass{lc-pink}{2} \end{bmatrix}}_B
BB
[47]ours\htmlClass{lc-blue}{\begin{bmatrix} 4 \\ 7 \end{bmatrix}}_{\text{ours}}
C−1C^{-1}
[13]C\htmlClass{lc-blue}{\begin{bmatrix} \htmlClass{lc-violet}{1} \\ \htmlClass{lc-violet}{3} \end{bmatrix}}_C
PC←B=C−1BP_{C \leftarrow B} = C^{-1}B

For any arrow in the plane, the arrow itself stays in the same place while its coordinate numbers change across Basis BB, the standard grid, and Basis CC. Multiplying [v]B[\arrowv{v}]_B by BB and then C−1C^{-1} gives the exact CC-coordinates of that same arrow.

The same arrow in space has a different step count on each grid, and the two matrix multiplications translate its coordinates from B's grid through our standard grid into C's grid.
3

Finding the Change-of-Basis Matrix by Row Reduction

To translate [v]B=[32]B[\arrowv{v}]_B = \arrowv{\begin{bmatrix} \bone{3} \\ \btwo{2} \end{bmatrix}}_B into Basis CC, we need the entries of PC←B=C−1BP_{C \leftarrow B} = C^{-1}B. Rather than finding C−1C^{-1} and multiplying two matrices, remember that multiplying C−1C^{-1} by B=[b1b2]B = \begin{bmatrix} \bone{b_1} & \btwo{b_2} \end{bmatrix} means multiplying C−1C^{-1} by one column of BB at a time:

PC←B=C−1[b1b2]=[C−1b1C−1b2]=[[b1]C[b2]C]\displaystyle \begin{aligned} P_{C \leftarrow B} &= C^{-1}\begin{bmatrix} \bone{b_1} & \btwo{b_2} \end{bmatrix} \\ &= \begin{bmatrix} C^{-1}\bone{b_1} & C^{-1}\btwo{b_2} \end{bmatrix} \\ &= \begin{bmatrix} \bone{[b_1]_C} & \btwo{[b_2]_C} \end{bmatrix} \end{aligned}

Because the columns of PC←BP_{C \leftarrow B} are just b1\bone{b_1} and b2\btwo{b_2} written in CC-coordinates, we can find both columns at once with a single row reduction.

Finding the Change-of-Basis Matrix by Row Reduction
[C∣B]=[112−11212]PC←B=  ?\displaystyle \begin{aligned} [C \mid B] &= \left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{1} & \bone{2} & \btwo{-1} \\ \cbasis{1} & \cbasis{2} & \bone{1} & \btwo{2} \end{array}\right] \\[0.6em] P_{C \leftarrow B} &= \;? \end{aligned}
Text solution
Set up the augmented matrix [C∣B][C \mid B]Put the target basis CC on the left and the starting basis BB on the right to solve for [b1]C\bone{[b_1]_C} and [b2]C\btwo{[b_2]_C} at the same time.
[C∣B]=[112−11212]\displaystyle [C \mid B] = \left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{1} & \bone{2} & \btwo{-1} \\ \cbasis{1} & \cbasis{2} & \bone{1} & \btwo{2} \end{array}\right]
Row-reduce the left side to the identity matrix IISubtract Row 1 from Row 2 to turn the bottom-left 1\cbasis{1} into 0\cbasis{0}, then subtract the new Row 2 from Row 1 to turn the top-right 1\cbasis{1} into 0\cbasis{0}, turning the right side into PC←BP_{C \leftarrow B}.
[112−11212]→R2−R1→R2[112−101−13]→R1−R2→R1[103−401−13]  ⟹  PC←B=[3−4−13]\displaystyle \begin{aligned} &\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{1} & \bone{2} & \btwo{-1} \\ \cbasis{1} & \cbasis{2} & \bone{1} & \btwo{2} \end{array}\right] \\ \xrightarrow{R_2 - R_1 \to R_2} &\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{1} & \bone{2} & \btwo{-1} \\ \cbasis{0} & \cbasis{1} & \bone{-1} & \btwo{3} \end{array}\right] \\ \xrightarrow{R_1 - R_2 \to R_1} &\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{0} & \bone{3} & \btwo{-4} \\ \cbasis{0} & \cbasis{1} & \bone{-1} & \btwo{3} \end{array}\right] \\ \implies &P_{C \leftarrow B} = \begin{bmatrix} \bone{3} & \btwo{-4} \\ \bone{-1} & \btwo{3} \end{bmatrix} \end{aligned}
Multiply by [v]B[\arrowv{v}]_B to translate the coordinates into Basis CCMultiply PC←BP_{C \leftarrow B} by [32]B\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{2} \end{bmatrix}}_B to find the CC-coordinates of the vector.
[v]C=PC←B[v]B=[3−4−13][32]B=[3(3)+(−4)(2)−1(3)+3(2)]C=[13]C\displaystyle \begin{aligned} [\arrowv{v}]_C &= P_{C \leftarrow B}[\arrowv{v}]_B \\ &= \begin{bmatrix} \bone{3} & \btwo{-4} \\ \bone{-1} & \btwo{3} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{2} \end{bmatrix}}_B \\ &= \arrowv{\begin{bmatrix} \bone{3(3)} + \btwo{(-4)(2)} \\ \bone{-1(3)} + \btwo{3(2)} \end{bmatrix}}_C \\ &= \arrowv{\begin{bmatrix} \cbasis{1} \\ \cbasis{3} \end{bmatrix}}_C \end{aligned}

So [13]C\arrowv{\begin{bmatrix} \cbasis{1} \\ \cbasis{3} \end{bmatrix}}_C in Basis CC and [32]B\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{2} \end{bmatrix}}_B in Basis BB both describe the exact same arrow [47]ours\arrowv{\begin{bmatrix} 4 \\ 7 \end{bmatrix}}_{\text{ours}} in our standard grid.

If a problem already gives b1\bone{b_1} and b2\btwo{b_2} as combinations of c1\cbasis{c_1} and c2\cbasis{c_2} (such as b1=3c1−c2\bone{b_1} = \bone{3}\cbasis{c_1} - \cbasis{c_2} and b2=−4c1+3c2\btwo{b_2} = \btwo{-4}\cbasis{c_1} + \btwo{3}\cbasis{c_2}), you can skip the row reduction completely and put those numbers straight into the first and second columns of PC←BP_{C \leftarrow B}.

4

Going Back the Other Way with the Inverse Matrix

Now suppose we want to translate in the opposite direction, from Basis CC back to Basis BB, using PB←CP_{B \leftarrow C}. Starting from scratch, we could row-reduce [B∣C]→[I∣PB←C][B \mid C] \to [I \mid P_{B \leftarrow C}], but since we already have PC←BP_{C \leftarrow B}, we can just invert it.

[abcd]−1=1ad−bc[d−b−ca]\displaystyle \begin{bmatrix} a & b \\ c & d \end{bmatrix}^{-1} = \frac{1}{ad - bc}\begin{bmatrix} d & -b \\ -c & a \end{bmatrix}
Going Back the Other Way with the Inverse Matrix
C-coordinates\text{C-coordinates}
B-coordinates\text{B-coordinates}
PC←B=[3−4−13]P_{C \leftarrow B} = \begin{bmatrix} \htmlClass{lc-green}{3} & \htmlClass{lc-pink}{-4} \\ \htmlClass{lc-green}{-1} & \htmlClass{lc-pink}{3} \end{bmatrix}
PB←C=  ?P_{B \leftarrow C} = \;?
Text solution
Invert PC←BP_{C \leftarrow B} to reverse the translation directionUse the 2×22 \times 2 inverse rule on PC←B=[3−4−13]P_{C \leftarrow B} = \begin{bmatrix} \bone{3} & \btwo{-4} \\ \bone{-1} & \btwo{3} \end{bmatrix} to find PB←CP_{B \leftarrow C}.
PB←C=(PC←B)−1=[3−4−13]−1=13(3)−(−4)(−1)[3413]=15[3413]\displaystyle \begin{aligned} P_{B \leftarrow C} &= (P_{C \leftarrow B})^{-1} \\ &= \begin{bmatrix} \bone{3} & \btwo{-4} \\ \bone{-1} & \btwo{3} \end{bmatrix}^{-1} \\ &= \frac{1}{3(3) - (-4)(-1)}\begin{bmatrix} 3 & 4 \\ 1 & 3 \end{bmatrix} \\ &= \frac{1}{5}\begin{bmatrix} 3 & 4 \\ 1 & 3 \end{bmatrix} \end{aligned}
Multiply by [v]C[\arrowv{v}]_C to get the BB-coordinates backMultiply PB←CP_{B \leftarrow C} by [13]C\arrowv{\begin{bmatrix} \cbasis{1} \\ \cbasis{3} \end{bmatrix}}_C to translate the vector back into Basis BB.
[v]B=PB←C[v]C=15[3413][13]C=15[1510]B=[32]B\displaystyle \begin{aligned} [\arrowv{v}]_B = P_{B \leftarrow C}[\arrowv{v}]_C &= \frac{1}{5}\begin{bmatrix} 3 & 4 \\ 1 & 3 \end{bmatrix}\arrowv{\begin{bmatrix} \cbasis{1} \\ \cbasis{3} \end{bmatrix}}_C \\ &= \frac{1}{5}\arrowv{\begin{bmatrix} \bone{15} \\ \btwo{10} \end{bmatrix}}_B \\ &= \arrowv{\begin{bmatrix} \bone{3} \\ \btwo{2} \end{bmatrix}}_B \end{aligned}

This brings us back to the original coordinate vector [32]B\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{2} \end{bmatrix}}_B we started with.

5

Simplifying Transformations by Switching Bases

So far we have only changed the coordinates of a fixed vector, but switching grids also helps us do geometric transformations that are hard to calculate in our standard (x,y)(x, y) grid, such as reflecting the vector [34]ours\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} across the tilted line through [21]\bone{\begin{bmatrix} 2 \\ 1 \end{bmatrix}}.

b1\htmlClass{lc-green}{b_1}
b2\htmlClass{lc-pink}{b_2}
[34]ours\htmlClass{lc-blue}{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}}
[50]ours\htmlClass{lc-blue}{\begin{bmatrix} 5 \\ 0 \end{bmatrix}}_{\text{ours}}
Simplifying Transformations by Switching Bases
[21]\htmlClass{lc-green}{\begin{bmatrix} 2 \\ 1 \end{bmatrix}}
[34]ours\htmlClass{lc-blue}{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}}
??
Text solution
Translate the input vector from the standard grid to Basis BBMultiply [34]ours\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} by B−1B^{-1} to split it into components along the line (b1\bone{b_1}) and perpendicular to the line (b2\btwo{b_2}).
B−1[34]ours=[21]B\displaystyle B^{-1}\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} = \arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B
Reflect the vector in Basis BBMultiply by the reflection matrix D=[100−1]D = \begin{bmatrix} \bone{1} & \btwo{0} \\ \bone{0} & \btwo{-1} \end{bmatrix} in Basis BB to keep the b1\bone{b_1} component along the line unchanged and flip the sign of the perpendicular b2\btwo{b_2} component.
D[21]B=[100−1][21]B=[2−1]B\displaystyle D\arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B = \begin{bmatrix} \bone{1} & \btwo{0} \\ \bone{0} & \btwo{-1} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B = \arrowv{\begin{bmatrix} \bone{2} \\ \btwo{-1} \end{bmatrix}}_B
Translate the reflected vector back to the standard gridMultiply by BB to convert the reflected BB-coordinates back into standard coordinates.
B[2−1]B=[2−112][2−1]B=2[21]−1[−12]=[50]ours\displaystyle \begin{aligned} B\arrowv{\begin{bmatrix} \bone{2} \\ \btwo{-1} \end{bmatrix}}_B &= \begin{bmatrix} \bone{2} & \btwo{-1} \\ \bone{1} & \btwo{2} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{2} \\ \btwo{-1} \end{bmatrix}}_B \\ &= \bone{2}\bone{\begin{bmatrix} 2 \\ 1 \end{bmatrix}} - \btwo{1}\btwo{\begin{bmatrix} -1 \\ 2 \end{bmatrix}} \\ &= \arrowv{\begin{bmatrix} 5 \\ 0 \end{bmatrix}}_{\text{ours}} \end{aligned}

So reflecting [34]ours\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} across the line through [21]\bone{\begin{bmatrix} 2 \\ 1 \end{bmatrix}} lands on [50]ours\arrowv{\begin{bmatrix} 5 \\ 0 \end{bmatrix}}_{\text{ours}}. Putting those three steps together from right to left connects the standard transformation matrix AA and the Basis-BB transformation matrix DD:

A=B⏟B tostandard D⏟transformin B B−1⏟standardto BandD=B−1⏟standardto B A⏟transformin standard B⏟B tostandard\displaystyle \begin{aligned} A &= \underbrace{B}_{\substack{B \text{ to} \\ \text{standard}}} \, \underbrace{D}_{\substack{\text{transform} \\ \text{in } B}} \, \underbrace{B^{-1}}_{\substack{\text{standard} \\ \text{to } B}} \qquad \text{and} \\ D &= \underbrace{B^{-1}}_{\substack{\text{standard} \\ \text{to } B}} \, \underbrace{A}_{\substack{\text{transform} \\ \text{in standard}}} \, \underbrace{B}_{\substack{B \text{ to} \\ \text{standard}}} \end{aligned}

Reflecting across the tilted line through b1=[21]\bone{b_1} = \bone{\begin{bmatrix} 2 \\ 1 \end{bmatrix}} mixes a vector's standard xx and yy coordinates together, but in Basis BB every vector's b1\bone{b_1} component along the line never changes and its b2\btwo{b_2} component perpendicular to the line simply flips sign.

Reflecting across a tilted line changes both regular coordinates at once, but in the tilted grid aligned with the line it just keeps the step along the line and flips the step across it.

Key concepts

Every change-of-basis problem comes down to one question: which grid are you starting in, and which grid are you going to?

When one of the grids is already the standard (x,y)(x, y) grid, you don't need a two-basis matrix PC←BP_{C \leftarrow B}. You only need the basis matrix B=[b1b2]B = \begin{bmatrix} \bone{b_1} & \btwo{b_2} \end{bmatrix} whose columns are the basis vectors. Multiplying a BB-vector by BB gives the vector in the standard grid, like B[31]B=[2−112][31]B=[55]oursB\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{1} \end{bmatrix}}_B = \begin{bmatrix} \bone{2} & \btwo{-1} \\ \bone{1} & \btwo{2} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{3} \\ \btwo{1} \end{bmatrix}}_B = \arrowv{\begin{bmatrix} 5 \\ 5 \end{bmatrix}}_{\text{ours}}, and multiplying a standard vector by B−1B^{-1} (or row-reducing [B∣v][B \mid \arrowv{v}]) translates it back into Basis BB, like B−1[34]ours=[21]BB^{-1}\arrowv{\begin{bmatrix} 3 \\ 4 \end{bmatrix}}_{\text{ours}} = \arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B.

When translating between two custom bases BB and CC, go through the standard grid with PC←B=C−1BP_{C \leftarrow B} = C^{-1}B so that [v]C=PC←B[v]B[\arrowv{v}]_C = P_{C \leftarrow B}[\arrowv{v}]_B. If BB and CC are given as standard vectors, put the target basis CC on the left (matching the arrow in C←BC \leftarrow B) and row-reduce [C∣B]→[I∣PC←B][C \mid B] \to [I \mid P_{C \leftarrow B}], as in [112−11212]→[103−401−13]\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{1} & \bone{2} & \btwo{-1} \\ \cbasis{1} & \cbasis{2} & \bone{1} & \btwo{2} \end{array}\right] \to \left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{0} & \bone{3} & \btwo{-4} \\ \cbasis{0} & \cbasis{1} & \bone{-1} & \btwo{3} \end{array}\right]. Skip row reduction whenever the BB-vectors are already written as combinations of the CC-vectors: if b1=3c1−c2\bone{b_1} = \bone{3}\cbasis{c_1} - \cbasis{c_2} and b2=−4c1+3c2\btwo{b_2} = \btwo{-4}\cbasis{c_1} + \btwo{3}\cbasis{c_2}, those coefficients are already [b1]C\bone{[b_1]_C} and [b2]C\btwo{[b_2]_C}, so they go straight into the columns of PC←B=[3−4−13]P_{C \leftarrow B} = \begin{bmatrix} \bone{3} & \btwo{-4} \\ \bone{-1} & \btwo{3} \end{bmatrix}. And to go back the other way, invert the matrix you already have: PB←C=(PC←B)−1P_{B \leftarrow C} = (P_{C \leftarrow B})^{-1}.

To switch bases for a transformation, match the rightmost matrix in the product to the basis of the input vector. To apply a simple Basis-BB matrix like D=[100−1]D = \begin{bmatrix} \bone{1} & \btwo{0} \\ \bone{0} & \btwo{-1} \end{bmatrix} to vectors in the standard grid, translate from standard to BB first with A=BDB−1A = BDB^{-1}; to see what a standard matrix AA looks like in Basis BB, translate from BB to standard first with D=B−1ABD = B^{-1}AB.

6

Practice

1

Basis BB consists of b1=[12]\bone{b_1} = \bone{\begin{bmatrix} 1 \\ 2 \end{bmatrix}} and b2=[35]\btwo{b_2} = \btwo{\begin{bmatrix} 3 \\ 5 \end{bmatrix}}. Find the standard coordinates of [u]B=[2−1]B[\arrowv{u}]_B = \begin{bmatrix} 2 \\ -1 \end{bmatrix}_B, and find the BB-coordinates of the standard vector w=[59]\arrowv{w} = \begin{bmatrix} 5 \\ 9 \end{bmatrix}.

Which vector is already a recipe of b1\bone{b_1} and b2\btwo{b_2} steps (multiply by BB), and which one starts in the standard grid and needs to be converted into Basis BB (multiply by B−1B^{-1})?

Put b1\bone{b_1} and b2\btwo{b_2} into the columns of B=[1325]B = \begin{bmatrix} \bone{1} & \btwo{3} \\ \bone{2} & \btwo{5} \end{bmatrix} (be careful not to put them into the rows as [1235]\begin{bmatrix} 1 & 2 \\ 3 & 5 \end{bmatrix}, which would give [01]\begin{bmatrix} 0 \\ 1 \end{bmatrix}) and multiply by [u]B[\arrowv{u}]_B:
u=B[u]B=[1325][2−1]B=[−1−1]\displaystyle \arrowv{u} = B[\arrowv{u}]_B = \begin{bmatrix} \bone{1} & \btwo{3} \\ \bone{2} & \btwo{5} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{2} \\ \btwo{-1} \end{bmatrix}}_B = \arrowv{\begin{bmatrix} -1 \\ -1 \end{bmatrix}}
To find [w]B[\arrowv{w}]_B, invert BB using the determinant 1(5)−3(2)=−11(5) - 3(2) = -1 (watch the negative sign on the determinant so you don't miss flipping the signs of B−1B^{-1} and get [−2−1]B\begin{bmatrix} -2 \\ -1 \end{bmatrix}_B, and don't multiply w\arrowv{w} by BB to get [3255]\begin{bmatrix} 32 \\ 55 \end{bmatrix}):
[w]B=B−1w=[−532−1][59]=[21]B\displaystyle [\arrowv{w}]_B = B^{-1}\arrowv{w} = \begin{bmatrix} -5 & 3 \\ 2 & -1 \end{bmatrix}\arrowv{\begin{bmatrix} 5 \\ 9 \end{bmatrix}} = \arrowv{\begin{bmatrix} \bone{2} \\ \btwo{1} \end{bmatrix}}_B
2

Basis BB has vectors b1=[31],b2=[−12]\bone{b_1} = \bone{\begin{bmatrix} 3 \\ 1 \end{bmatrix}}, \btwo{b_2} = \btwo{\begin{bmatrix} -1 \\ 2 \end{bmatrix}} and Basis CC has vectors c1=[11],c2=[23]\cbasis{c_1} = \cbasis{\begin{bmatrix} 1 \\ 1 \end{bmatrix}}, \cbasis{c_2} = \cbasis{\begin{bmatrix} 2 \\ 3 \end{bmatrix}}. Row-reduce [C∣B][C \mid B] to find the change-of-basis matrix PC←BP_{C \leftarrow B}, then use it to find [v]C[\arrowv{v}]_C for [v]B=[12]B[\arrowv{v}]_B = \begin{bmatrix} 1 \\ 2 \end{bmatrix}_B.

Check the arrow in PC←BP_{C \leftarrow B}: which basis matrix goes on the left side of the augmented line so that row reduction turns it into II?

Put the target basis CC on the left and BB on the right (setting up [B∣C][B \mid C] in the order the bases are mentioned in the problem instead of [C∣B][C \mid B] would find PB←C=17[3727]P_{B \leftarrow C} = \frac{1}{7}\begin{bmatrix} 3 & 7 \\ 2 & 7 \end{bmatrix} and give 17[1716]\frac{1}{7}\begin{bmatrix} 17 \\ 16 \end{bmatrix}). When turning the top-right 2\cbasis{2} into 0\cbasis{0} with R1−2R2→R1R_1 - 2R_2 \to R_1, watch the signs in 3−2(−2)=7\bone{3} - 2(\bone{-2}) = \bone{7} and −1−2(3)=−7\btwo{-1} - 2(\btwo{3}) = \btwo{-7} (subtracting the wrong sign gives [−15−23]\begin{bmatrix} -1 & 5 \\ -2 & 3 \end{bmatrix} and [94]C\begin{bmatrix} 9 \\ 4 \end{bmatrix}_C):
[123−11312]→R2−R1→R2[123−101−23]→R1−2R2→R1[107−701−23]  ⟹  PC←B=[7−7−23]\displaystyle \begin{aligned} &\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{2} & \bone{3} & \btwo{-1} \\ \cbasis{1} & \cbasis{3} & \bone{1} & \btwo{2} \end{array}\right] \\ \xrightarrow{R_2 - R_1 \to R_2} &\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{2} & \bone{3} & \btwo{-1} \\ \cbasis{0} & \cbasis{1} & \bone{-2} & \btwo{3} \end{array}\right] \\ \xrightarrow{R_1 - 2R_2 \to R_1} &\left[\begin{array}{cc|cc} \cbasis{1} & \cbasis{0} & \bone{7} & \btwo{-7} \\ \cbasis{0} & \cbasis{1} & \bone{-2} & \btwo{3} \end{array}\right] \\ \implies &P_{C \leftarrow B} = \begin{bmatrix} \bone{7} & \btwo{-7} \\ \bone{-2} & \btwo{3} \end{bmatrix} \end{aligned}
Multiply PC←BP_{C \leftarrow B} by [v]B[\arrowv{v}]_B:
[v]C=[7−7−23][12]B=[−74]C\displaystyle [\arrowv{v}]_C = \begin{bmatrix} \bone{7} & \btwo{-7} \\ \bone{-2} & \btwo{3} \end{bmatrix}\arrowv{\begin{bmatrix} \bone{1} \\ \btwo{2} \end{bmatrix}}_B = \arrowv{\begin{bmatrix} \cbasis{-7} \\ \cbasis{4} \end{bmatrix}}_C
3

Two bases B={b1,b2}B = \{\bone{b_1}, \btwo{b_2}\} and C={c1,c2}C = \{\cbasis{c_1}, \cbasis{c_2}\} satisfy b1=4c1−c2\bone{b_1} = 4\cbasis{c_1} - \cbasis{c_2} and b2=−3c1+2c2\btwo{b_2} = -3\cbasis{c_1} + 2\cbasis{c_2}. Write down PC←BP_{C \leftarrow B} directly from these equations, then find PB←CP_{B \leftarrow C} and use it to find [v]B[\arrowv{v}]_B when [v]C=[16]C[\arrowv{v}]_C = \begin{bmatrix} 1 \\ 6 \end{bmatrix}_C.

The given equations already tell you [b1]C\bone{[b_1]_C} and [b2]C\btwo{[b_2]_C}. Do those coordinate pairs go into the rows or the columns of PC←BP_{C \leftarrow B}?

Put the CC-coordinates [b1]C=[4−1]C\bone{[b_1]_C} = \bone{\begin{bmatrix} 4 \\ -1 \end{bmatrix}}_C and [b2]C=[−32]C\btwo{[b_2]_C} = \btwo{\begin{bmatrix} -3 \\ 2 \end{bmatrix}}_C into the columns of PC←BP_{C \leftarrow B} (writing them across the rows as [4−1−32]\begin{bmatrix} 4 & -1 \\ -3 & 2 \end{bmatrix} transposes the matrix and gives 15[827]B\frac{1}{5}\begin{bmatrix} 8 \\ 27 \end{bmatrix}_B):
PC←B=[4−3−12]\displaystyle P_{C \leftarrow B} = \begin{bmatrix} \bone{4} & \btwo{-3} \\ \bone{-1} & \btwo{2} \end{bmatrix}
Invert PC←BP_{C \leftarrow B} to get PB←CP_{B \leftarrow C} before multiplying by [v]C[\arrowv{v}]_C (multiplying [16]C\arrowv{\begin{bmatrix} \cbasis{1} \\ \cbasis{6} \end{bmatrix}}_C directly by PC←BP_{C \leftarrow B} without inverting first translates in the wrong direction and gives [−1411]\begin{bmatrix} -14 \\ 11 \end{bmatrix}):
PB←C=(PC←B)−1=15[2314]\displaystyle P_{B \leftarrow C} = (P_{C \leftarrow B})^{-1} = \frac{1}{5}\begin{bmatrix} 2 & 3 \\ 1 & 4 \end{bmatrix}
[v]B=15[2314][16]C=[45]B\displaystyle [\arrowv{v}]_B = \frac{1}{5}\begin{bmatrix} 2 & 3 \\ 1 & 4 \end{bmatrix}\arrowv{\begin{bmatrix} \cbasis{1} \\ \cbasis{6} \end{bmatrix}}_C = \arrowv{\begin{bmatrix} \bone{4} \\ \btwo{5} \end{bmatrix}}_B
4

Using the basis b1=[31]\bone{b_1} = \bone{\begin{bmatrix} 3 \\ 1 \end{bmatrix}} (along the line) and b2=[−13]\btwo{b_2} = \btwo{\begin{bmatrix} -1 \\ 3 \end{bmatrix}} (perpendicular to the line), find the standard matrix A=BDB−1A = BDB^{-1} that reflects vectors across the line through the origin and [31]\bone{\begin{bmatrix} 3 \\ 1 \end{bmatrix}}, and use it to reflect the standard vector [17]\begin{bmatrix} 1 \\ 7 \end{bmatrix}.

In Basis BB, which coordinate stays the same and which coordinate flips sign when you reflect across the line along b1\bone{b_1}?

In Basis BB, reflection across b1\bone{b_1} keeps the first coordinate and flips the second, so D=[100−1]D = \begin{bmatrix} \bone{1} & \btwo{0} \\ \bone{0} & \btwo{-1} \end{bmatrix} (flipping the first coordinate with [−1001]\begin{bmatrix} -1 & 0 \\ 0 & 1 \end{bmatrix} reflects across the perpendicular vector b2\btwo{b_2} instead and gives [−55]\begin{bmatrix} -5 \\ 5 \end{bmatrix}). With B=[3−113]B = \begin{bmatrix} \bone{3} & \btwo{-1} \\ \bone{1} & \btwo{3} \end{bmatrix} and B−1=110[31−13]B^{-1} = \frac{1}{10}\begin{bmatrix} 3 & 1 \\ -1 & 3 \end{bmatrix}, multiply in the order BDB−1BDB^{-1} (reversing the order to B−1DBB^{-1}DB flips the signs off the diagonal to 15[4−3−3−4]\frac{1}{5}\begin{bmatrix} 4 & -3 \\ -3 & -4 \end{bmatrix} and gives 15[−17−31]\frac{1}{5}\begin{bmatrix} -17 \\ -31 \end{bmatrix}):
A=BDB−1=[3−113][100−1](110[31−13])=15[433−4]\displaystyle \begin{aligned} A &= BDB^{-1} \\ &= \begin{bmatrix} \bone{3} & \btwo{-1} \\ \bone{1} & \btwo{3} \end{bmatrix}\begin{bmatrix} \bone{1} & \btwo{0} \\ \bone{0} & \btwo{-1} \end{bmatrix} \\ &\qquad \left(\frac{1}{10}\begin{bmatrix} 3 & 1 \\ -1 & 3 \end{bmatrix}\right) \\ &= \frac{1}{5}\begin{bmatrix} 4 & 3 \\ 3 & -4 \end{bmatrix} \end{aligned}
A[17]=15[433−4][17]=[5−5]\displaystyle A\arrowv{\begin{bmatrix} 1 \\ 7 \end{bmatrix}} = \frac{1}{5}\begin{bmatrix} 4 & 3 \\ 3 & -4 \end{bmatrix}\arrowv{\begin{bmatrix} 1 \\ 7 \end{bmatrix}} = \arrowv{\begin{bmatrix} 5 \\ -5 \end{bmatrix}}
5

A linear transformation in the standard grid is given by A=[4−23−1]A = \begin{bmatrix} 4 & -2 \\ 3 & -1 \end{bmatrix}. Find the matrix D=B−1ABD = B^{-1}AB that represents this transformation in Basis BB with b1=[11]\bone{b_1} = \bone{\begin{bmatrix} 1 \\ 1 \end{bmatrix}} and b2=[23]\btwo{b_2} = \btwo{\begin{bmatrix} 2 \\ 3 \end{bmatrix}}.

Follow an input vector starting in Basis BB: which matrix on the right translates it into the standard grid first so that AA can multiply it?

With B=[1213]B = \begin{bmatrix} \bone{1} & \btwo{2} \\ \bone{1} & \btwo{3} \end{bmatrix}, the 2×22 \times 2 inverse rule gives B−1=[3−2−11]B^{-1} = \begin{bmatrix} 3 & -2 \\ -1 & 1 \end{bmatrix}. Put BB on the right and B−1B^{-1} on the left to form D=B−1ABD = B^{-1}AB (putting B−1B^{-1} on the right as BAB−1BAB^{-1} translates in the wrong direction and gives [34−2444−31]\begin{bmatrix} 34 & -24 \\ 44 & -31 \end{bmatrix}):
D=B−1AB=[3−2−11][4−23−1][1213]=[3−2−11][2223]=[2001]\displaystyle \begin{aligned} D &= B^{-1}AB \\ &= \begin{bmatrix} 3 & -2 \\ -1 & 1 \end{bmatrix}\begin{bmatrix} 4 & -2 \\ 3 & -1 \end{bmatrix}\begin{bmatrix} \bone{1} & \btwo{2} \\ \bone{1} & \btwo{3} \end{bmatrix} \\ &= \begin{bmatrix} 3 & -2 \\ -1 & 1 \end{bmatrix}\begin{bmatrix} \bone{2} & \btwo{2} \\ \bone{2} & \btwo{3} \end{bmatrix} \\ &= \begin{bmatrix} \bone{2} & \btwo{0} \\ \bone{0} & \btwo{1} \end{bmatrix} \end{aligned}