Class 12 Maths, Worked in Full Past papersGet updates
Chapter 3 · Matrices

Example 23: Powers of the rotation matrix, by induction

ExamplesMiscellaneous examples · Chapter 3
The question

If

A=[cos⁡θsin⁡θ−sin⁡θcos⁡θ],A = \begin{bmatrix} \cos\theta & \sin\theta \\ -\sin\theta & \cos\theta \end{bmatrix},

then prove that

An=[cos⁡nθsin⁡nθ−sin⁡nθcos⁡nθ],n∈N.A^n = \begin{bmatrix} \cos n\theta & \sin n\theta \\ -\sin n\theta & \cos n\theta \end{bmatrix}, \quad n \in \N.
Exam answer

Let P(n)P(n) be the statement

An=[cos⁡nθsin⁡nθ−sin⁡nθcos⁡nθ]A^n = \begin{bmatrix} \cos n\theta & \sin n\theta \\ -\sin n\theta & \cos n\theta \end{bmatrix}

For n=1n = 1⁠, the right side is AA itself and A1=AA^1 = A⁠, so P(1)P(1) is true.

Assume P(k)P(k) is true for some k∈Nk \in \N⁠:

Ak=[cos⁡kθsin⁡kθ−sin⁡kθcos⁡kθ]A^k = \begin{bmatrix} \cos k\theta & \sin k\theta \\ -\sin k\theta & \cos k\theta \end{bmatrix}

Then Ak+1=A⋅AkA^{k+1} = A{\cdot} A^k⁠. Let Ak+1=[cij]A^{k+1} = [c_{ij}]⁠. Using the formulas for cos⁡(x+y)\cos(x + y) and sin⁡(x+y)\sin(x + y)⁠,

c11=cos⁡θcos⁡kθ−sin⁡θsin⁡kθ=cos⁡(k+1)θc12=cos⁡θsin⁡kθ+sin⁡θcos⁡kθ=sin⁡(k+1)θc21=−sin⁡θcos⁡kθ−cos⁡θsin⁡kθ=−sin⁡(k+1)θc22=−sin⁡θsin⁡kθ+cos⁡θcos⁡kθ=cos⁡(k+1)θ\begin{aligned} c_{11} &= \cos\theta\cos k\theta - \sin\theta\sin k\theta \\ &= \cos(k + 1)\theta \\ c_{12} &= \cos\theta\sin k\theta + \sin\theta\cos k\theta \\ &= \sin(k + 1)\theta \\ c_{21} &= -\sin\theta\cos k\theta - \cos\theta\sin k\theta \\ &= -\sin(k + 1)\theta \\ c_{22} &= -\sin\theta\sin k\theta + \cos\theta\cos k\theta \\ &= \cos(k + 1)\theta \end{aligned} ∴ Ak+1=[cos⁡(k+1)θsin⁡(k+1)θ−sin⁡(k+1)θcos⁡(k+1)θ]\begin{aligned} &\therefore\ A^{k+1} \\ &= \begin{bmatrix} \cos(k + 1)\theta & \sin(k + 1)\theta \\ -\sin(k + 1)\theta & \cos(k + 1)\theta \end{bmatrix} \end{aligned}

So P(k+1)P(k + 1) is true whenever P(k)P(k) is true.

Hence, by the principle of mathematical induction, P(n)P(n) is true for all n∈Nn \in \N⁠.

Want the reason behind every step, the basics and the checks?

The one idea this problem is built on

Multiplying by AA adds θ\theta to the angle. When AA multiplies a matrix of the same form with angle ϕ\phi⁠, the four entries of the product are exactly the four compound angle expansions of cos⁡(θ+ϕ)\cos(\theta+\phi) and sin⁡(θ+ϕ)\sin(\theta+\phi)⁠. So each extra factor of AA pushes the angle up by θ\theta⁠, and after nn factors the angle is nθn\theta⁠. Mathematical induction turns this "each step adds θ\theta" into a proof for every natural number nn⁠.

Given, to show, and the plan
Given
The 2×22\times2 matrix AA above, where θ\theta is any real number.
To show
For every n∈N={1,2,3,… }n \in \N = \{1, 2, 3, \dots\}⁠, AnA^n is the matrix with nθn\theta in place of θ\theta⁠.
Plan
Mathematical induction: state P(n)P(n)⁠; prove P(1)P(1)⁠; assume P(k)P(k) and prove P(k+1)P(k+1) by multiplying A⋅AkA \cdot A^k entry by entry and using the compound angle formulas; conclude.

Basics you need

1Mathematical induction

To prove a statement P(n)P(n) for every n∈Nn \in \N⁠: (a) prove P(1)P(1)⁠; (b) prove that whenever P(k)P(k) is true for some k∈Nk \in \N⁠, P(k+1)P(k+1) is also true. Then P(1)P(1) gives P(2)P(2)⁠, which gives P(3)P(3)⁠, and so on without end. Basics: compound angles and induction

2Compound angle formulas

For all real x,yx, y⁠:

cos⁡(x+y)=cos⁡xcos⁡y−sin⁡xsin⁡y,sin⁡(x+y)=sin⁡xcos⁡y+cos⁡xsin⁡y.\begin{aligned} \cos(x+y) &= \cos x\cos y - \sin x\sin y,\\ \sin(x+y) &= \sin x\cos y + \cos x\sin y. \end{aligned}

(Class XI, Trigonometric Functions.) Basics: compound angles and induction

3Multiplying 2×22\times2 matrices

The (i,j)(i,j) entry of XYXY is row ii of XX times column jj of YY⁠: xi1y1j+xi2y2jx_{i1}y_{1j} + x_{i2}y_{2j}⁠. Basics: pair up, multiply, add

4Powers of a square matrix

A1=AA^1 = A⁠, A2=AAA^2 = AA⁠, and Ak+1A^{k+1} is AkA^k multiplied by one more AA⁠. Because matrix multiplication is associative, it does not matter whether the extra AA goes on the left or the right: A⋅Ak=Ak⋅AA\cdot A^k = A^k \cdot A⁠, since both are a product of k+1k+1 copies of AA⁠.

Solution

  1. State the statement P(n)P(n)

    For each natural number nn⁠, let P(n)P(n) be the statement

    P(n):An=[cos⁡nθsin⁡nθ−sin⁡nθcos⁡nθ].P(n):\quad A^n = \begin{bmatrix} \cos n\theta & \sin n\theta \\ -\sin n\theta & \cos n\theta \end{bmatrix}.

    We must prove P(n)P(n) is true for every n∈Nn \in \N⁠, for the given (arbitrary) θ\theta⁠.

    Induction needs one statement for each nn⁠, written so we can plug in n=1n = 1 or n=kn = k or n=k+1n = k+1⁠. Writing P(n)P(n) out exactly also fixes what "the result for k+1k+1" means: the same matrix with (k+1)θ(k+1)\theta in every entry.
    Principle of mathematical induction (Class XI). Basics: compound angles and induction
  2. Base case: P(1)P(1) is true

    With n=1n = 1⁠, the right side of P(1)P(1) has 1⋅θ=θ1\cdot\theta = \theta in every entry:

    [cos⁡θsin⁡θ−sin⁡θcos⁡θ],\begin{bmatrix} \cos\theta & \sin\theta \\ -\sin\theta & \cos\theta \end{bmatrix},

    which is exactly AA⁠. Since A1=AA^1 = A⁠, P(1)P(1) is true.

    The base case is the first domino. It is easy here, but it must be stated: without it, the inductive step alone proves nothing.
  3. Inductive hypothesis: assume P(k)P(k)

    Let kk be any natural number for which P(k)P(k) is true, that is,

    Ak=[cos⁡kθsin⁡kθ−sin⁡kθcos⁡kθ].A^k = \begin{bmatrix} \cos k\theta & \sin k\theta \\ -\sin k\theta & \cos k\theta \end{bmatrix}.

    We now show that P(k+1)P(k+1) must then be true.

    We do not claim P(k)P(k) is true; we only ask "if it were, would P(k+1)P(k+1) follow?" Showing this for an arbitrary kk covers every link of the chain P(1)⇒P(2)⇒P(3)⇒…P(1) \Rightarrow P(2) \Rightarrow P(3) \Rightarrow \dots at once.
  4. Write Ak+1A^{k+1} as a product and check the orders

    Ak+1=A⋅AkA^{k+1} = A \cdot A^k (Basic 4). Both AA and AkA^k are 2×22\times2⁠: the columns of AA (2) match the rows of AkA^k (2), so the product is defined and is 2×22\times2⁠. Using the hypothesis from Step 3,

    Ak+1=[cos⁡θsin⁡θ−sin⁡θcos⁡θ][cos⁡kθsin⁡kθ−sin⁡kθcos⁡kθ].A^{k+1} = \begin{bmatrix} \cos\theta & \sin\theta \\ -\sin\theta & \cos\theta \end{bmatrix} \begin{bmatrix} \cos k\theta & \sin k\theta \\ -\sin k\theta & \cos k\theta \end{bmatrix}.
    This is where the hypothesis is used: it lets us replace AkA^k by an explicit matrix that we can multiply. Splitting off one factor of AA links the kk-th power to the (k+1)(k+1)-th power.
    Associative law, Section 3.4.6: (XY)Z=X(YZ)(XY)Z = X(YZ)⁠, which is why A⋅AkA \cdot A^k and Ak⋅AA^k\cdot A are both Ak+1A^{k+1}⁠. Lesson: properties of the operations
  5. Multiply, one entry at a time

    Each entry is (row of AA⁠) times (column of AkA^k⁠):

    c11=cos⁡θcos⁡kθ+sin⁡θ(−sin⁡kθ)=cos⁡θcos⁡kθ−sin⁡θsin⁡kθ,c12=cos⁡θsin⁡kθ+sin⁡θcos⁡kθ,c21=(−sin⁡θ)cos⁡kθ+cos⁡θ(−sin⁡kθ)=−(sin⁡θcos⁡kθ+cos⁡θsin⁡kθ),c22=(−sin⁡θ)sin⁡kθ+cos⁡θcos⁡kθ=cos⁡θcos⁡kθ−sin⁡θsin⁡kθ.\begin{aligned} c_{11} &= \cos\theta\cos k\theta + \sin\theta(-\sin k\theta)\\ &= \cos\theta\cos k\theta - \sin\theta\sin k\theta,\\[4pt] c_{12} &= \cos\theta\sin k\theta + \sin\theta\cos k\theta,\\[4pt] c_{21} &= (-\sin\theta)\cos k\theta + \cos\theta(-\sin k\theta)\\ &= -(\sin\theta\cos k\theta + \cos\theta\sin k\theta),\\[4pt] c_{22} &= (-\sin\theta)\sin k\theta + \cos\theta\cos k\theta\\ &= \cos\theta\cos k\theta - \sin\theta\sin k\theta. \end{aligned}
    Row 1 of AA is cos⁡θ,sin⁡θ\cos\theta, \sin\theta and row 2 is −sin⁡θ,cos⁡θ-\sin\theta, \cos\theta⁠; column 1 of AkA^k is cos⁡kθ,−sin⁡kθ\cos k\theta, -\sin k\theta and column 2 is sin⁡kθ,cos⁡kθ\sin k\theta, \cos k\theta⁠. In c21c_{21} both terms are negative, so we take out a minus sign; in c22c_{22} we just reorder the two terms. These rearrangements are chosen to make each entry match a compound angle formula exactly.
    Matrix multiplication, Section 3.4.5. Basics: pair up, multiply, add
  6. Recognise the compound angle formulas

    Put x=θx = \theta and y=kθy = k\theta in Basic 2:

    c11=cos⁡(θ+kθ)=cos⁡(k+1)θ,c12=sin⁡(θ+kθ)=sin⁡(k+1)θ,c21=−sin⁡(θ+kθ)=−sin⁡(k+1)θ,c22=cos⁡(θ+kθ)=cos⁡(k+1)θ.\begin{aligned} c_{11} &= \cos(\theta + k\theta) = \cos(k+1)\theta,\\ c_{12} &= \sin(\theta + k\theta) = \sin(k+1)\theta,\\ c_{21} &= -\sin(\theta + k\theta) = -\sin(k+1)\theta,\\ c_{22} &= \cos(\theta + k\theta) = \cos(k+1)\theta. \end{aligned}

    For c12c_{12}⁠, note that cos⁡θsin⁡kθ+sin⁡θcos⁡kθ\cos\theta\sin k\theta + \sin\theta\cos k\theta is sin⁡θcos⁡kθ+cos⁡θsin⁡kθ\sin\theta\cos k\theta + \cos\theta\sin k\theta with the two terms swapped, which is sin⁡(θ+kθ)\sin(\theta + k\theta)⁠. Also θ+kθ=(k+1)θ\theta + k\theta = (k+1)\theta⁠. So

    Ak+1=[cos⁡(k+1)θsin⁡(k+1)θ−sin⁡(k+1)θcos⁡(k+1)θ],A^{k+1} = \begin{bmatrix} \cos(k+1)\theta & \sin(k+1)\theta \\ -\sin(k+1)\theta & \cos(k+1)\theta \end{bmatrix},

    which is exactly P(k+1)P(k+1)⁠.

    The compound angle formulas hold for all real angles, so no restriction on θ\theta or kk is needed. This is the moment the idea of the problem becomes visible: multiplying by AA turned the angle kθk\theta into kθ+θk\theta + \theta⁠.
    Compound angle formulas (Class XI, Chapter 3). Basics: compound angles and induction
  7. Conclude by induction

    P(1)P(1) is true (Step 2), and for every k∈Nk \in \N⁠, P(k)P(k) true implies P(k+1)P(k+1) true (Steps 3 to 6). By the principle of mathematical induction, P(n)P(n) is true for every natural number nn⁠:

    An=[cos⁡nθsin⁡nθ−sin⁡nθcos⁡nθ],n∈N.A^n = \begin{bmatrix} \cos n\theta & \sin n\theta \\ -\sin n\theta & \cos n\theta \end{bmatrix}, \quad n \in \N.
    (1, 0) A·(1, 0) A²·(1, 0) A³·(1, 0) θ θ θ θ = 30° in the picture each factor A turns clockwise by θ, so Aⁿ turns by nθ
    What AA does. Acting on the column with entries 1,01, 0⁠, the matrix AA gives the column cos⁡θ,−sin⁡θ\cos\theta, -\sin\theta⁠: the point is turned clockwise through θ\theta on the unit circle. Each further factor of AA turns it through another θ\theta⁠, so AnA^n turns it through nθn\theta⁠.
    Both conditions of induction have been checked, so the chain of implications reaches every natural number. The picture shows the same fact geometrically: AA is a rotation through θ\theta (clockwise, because of where the minus sign sits), and doing it nn times is a rotation through nθn\theta⁠.
Result

For every natural number nn⁠,

An=[cos⁡nθsin⁡nθ−sin⁡nθcos⁡nθ],A^n = \begin{bmatrix} \cos n\theta & \sin n\theta \\ -\sin n\theta & \cos n\theta \end{bmatrix},

proved by mathematical induction.

Check it with numbers

Take θ=0.3\theta = 0.3 radians and n=5n = 5⁠. Then cos⁡0.3≈0.955336\cos 0.3 \approx 0.955336 and sin⁡0.3≈0.295520\sin 0.3 \approx 0.295520⁠. Multiplying AA by itself five times numerically gives A5A^5 with entries 0.0707370.070737⁠, 0.9974950.997495⁠, −0.997495-0.997495⁠, 0.0707370.070737 (rows 1 and 2), and indeed cos⁡1.5≈0.070737\cos 1.5 \approx 0.070737 and sin⁡1.5≈0.997495\sin 1.5 \approx 0.997495⁠. For an exact case, take θ=π2\theta = \frac{\pi}{2}⁠, which gives the matrices below. The formula predicts that A4A^4 has cos⁡2π=1\cos 2\pi = 1 and sin⁡2π=0\sin 2\pi = 0⁠, so A4A^4 should be the identity matrix. Direct multiplication confirms it: the entries of A2=AAA^2 = AA are 0⋅0+1⋅(−1)=−10\cdot0 + 1\cdot(-1) = -1⁠, 0⋅1+1⋅0=00\cdot1 + 1\cdot0 = 0⁠, (−1)⋅0+0⋅(−1)=0(-1)\cdot0 + 0\cdot(-1) = 0 and (−1)⋅1+0⋅0=−1(-1)\cdot1 + 0\cdot0 = -1⁠, so A2=−IA^2 = -I⁠, and then A4=A2A2=(−I)(−I)=IA^4 = A^2A^2 = (-I)(-I) = I⁠.

θ=π2:A=[01−10],\theta = \tfrac{\pi}{2}:\quad A = \begin{bmatrix} 0 & 1 \\ -1 & 0 \end{bmatrix}, A2=[−100−1],A4=A2A2=I.A^2 = \begin{bmatrix} -1 & 0 \\ 0 & -1 \end{bmatrix},\quad A^4 = A^2A^2 = I.
Common mistakes
  1. Raising each entry to the power nn⁠. AnA^n is not the matrix of cos⁡nθ\cos^n\theta⁠, sin⁡nθ\sin^n\theta⁠, and so on. Powers of a matrix mean repeated matrix multiplication.
  2. Assuming what is to be proved. In the inductive step you may assume P(k)P(k) only, and must derive P(k+1)P(k+1)⁠.
  3. Losing the minus signs. In c21c_{21} both products are negative; taking out the minus sign is what reveals −sin⁡(k+1)θ-\sin(k+1)\theta⁠.
  4. Skipping the base case. The inductive step alone does not prove anything about n=1n = 1⁠, and so nothing about any nn⁠.
What to take away
  1. For a matrix whose powers follow a pattern, guess the pattern and prove it by induction with Ak+1=A⋅AkA^{k+1} = A\cdot A^k⁠.
  2. The entries of a product of two "rotation" matrices are precisely the compound angle formulas: multiplying the matrices adds the angles.
  3. The same computation shows more generally that the matrices with angles α\alpha and β\beta multiply to the one with angle α+β\alpha+\beta⁠.

Every fact used, and where it comes from

FactUsed inSource
Principle of mathematical inductionSteps 1 to 3, 7Class XI (Basics: induction)
Product of 2×22\times2 matrices, entry ruleSteps 4, 5Section 3.4.5 (Lesson: matrix multiplication)
Associative law, so A⋅Ak=Ak+1A\cdot A^k = A^{k+1}Step 4Section 3.4.6 (Lesson: properties)
cos⁡(x+y)\cos(x+y)⁠, sin⁡(x+y)\sin(x+y) expansionsStep 6Class XI, Trigonometric Functions
Equal matrices: same order, equal entriesSteps 2, 6Definition 2, Section 3.3.1
Method of this proofWhole pageMiscellaneous Examples, Example 23 (Lesson: miscellaneous ideas)
Found an error, or want more?

Tell us if something on this page is wrong or unclear, or if you would like a fuller explanation, another method or a new feature. Send feedback on this problem →