Menu
Sign up Sign in

AI HL · AHL 9

Differential Equations

AHL 9.8

Solving coupled linear systems exactly

AHL 9.8

Solving coupled linear systems exactly

Target: I can write a pair of linear equations dx/dt = ax + by, dy/dt = cx + dy in matrix form. (+4 more)

By the end of this section

Key points

Matrix form

  • A linear system \(x'=ax+by,\ y'=cx+dy\) is written \(\begin{pmatrix}x'\\y'\end{pmatrix}=\mathbf M\begin{pmatrix}x\\y\end{pmatrix}\) with \(\mathbf M=\begin{pmatrix}a&b\\c&d\end{pmatrix}\).
  • x' means dx/dt. Row 1 holds the coefficients in dx/dt, row 2 those in dy/dt. A missing term has coefficient 0.
  • Only a linear system can be solved exactly in this way.

Eigenvalues and eigenvectors

  • If \(\mathbf M\mathbf v=\lambda\mathbf v\) with \(\mathbf v\) not the zero vector, then \(\lambda\) is an eigenvalue and \(\mathbf v\) its eigenvector.
  • Eigenvalues solve \(\det(\mathbf M-\lambda\mathbf I)=0\), that is \(\lambda^2-(\text{trace})\lambda+\det\mathbf M=0\).
  • For each eigenvalue, solve \((\mathbf M-\lambda\mathbf I)\mathbf v=\mathbf 0\). One of the two equations is enough. Any non-zero multiple of v is an eigenvector.

The general solution

  • If M has real, distinct eigenvalues \(\lambda_1,\lambda_2\) with eigenvectors \(\mathbf p_1,\mathbf p_2\): \(\begin{pmatrix}x\\y\end{pmatrix}=Ae^{\lambda_1t}\mathbf p_1+Be^{\lambda_2t}\mathbf p_2\).
  • A and B are two arbitrary constants. Read x from the top row and y from the bottom row.
  • You do not need to prove this.

Initial conditions

  • Put t = 0: the exponentials become 1, so \(A\mathbf p_1+B\mathbf p_2=\begin{pmatrix}x(0)\\y(0)\end{pmatrix}\). Solve for A and B.
  • If x(0) and x'(0) are given, put t = 0 in the first equation, \(x'(0)=a\,x(0)+b\,y(0)\), to find y(0). Then continue as above.
  • Check the answer by putting t = 0 back in.

On the calculator

Eigenvalues and eigenvectors

  • Enter the matrix M of the system.
  • Use the calculator's eigenvalue and eigenvector commands. It gives the eigenvalues, and for each one a vector pointing the right way.
  • The calculator may scale the eigenvector (for example to length 1). Any non-zero multiple is correct: rescale to whole numbers.

Screenshot and video to come

Solving for A and B

  • At t = 0 the exponentials are 1, so x(0) and y(0) give two linear equations in A and B.
  • Use the simultaneous equation solver for A and B, or the matrix solver with the eigenvector matrix.
  • Substitute A and B back to check x(0) and y(0).

Screenshot and video to come

Worked examples

Problems like the ones in class. Pick one, try it, then open the working.
Matrix form and eigenvalues

Write the system \(\dfrac{dx}{dt}=2x+y\), \(\dfrac{dy}{dt}=3x+4y\) in the form \(\begin{pmatrix}x'\\y'\end{pmatrix}=\mathbf M\begin{pmatrix}x\\y\end{pmatrix}\), and find the eigenvalues of \(\mathbf M\).

Working
  1. Matrix form

    \(\begin{pmatrix}x'\\y'\end{pmatrix}=\begin{pmatrix}2&1\\3&4\end{pmatrix}\begin{pmatrix}x\\y\end{pmatrix}\)

    Row 1 holds the coefficients of x and y in dx/dt. Row 2 holds those in dy/dt. x' means dx/dt.

  2. Characteristic equation

    \(\det(\mathbf M-\lambda\mathbf I)=(2-\lambda)(4-\lambda)-(1)(3)=0\)

    Eigenvalues are the values of λ for which Mv = λv has a non-zero solution v.

  3. Tidy

    \(\lambda^2-6\lambda+5=0\)

    The 6 is the trace (2 + 4) and the 5 is the determinant (8 − 3).

  4. Solve

    \((\lambda-1)(\lambda-5)=0\Rightarrow\lambda=1\ \text{or}\ 5\)

    From the calculator: eigenvalues 1 and 5. Use it to check.

Video walkthrough coming
Eigenvectors and the general solution

The matrix \(\mathbf M=\begin{pmatrix}2&1\\3&4\end{pmatrix}\) has eigenvalues 1 and 5, from the system \(x'=2x+y,\ y'=3x+4y\). Find the eigenvector for each eigenvalue and write down the general solution.

Working
  1. Eigenvector for λ = 1

    \((\mathbf M-1\mathbf I)\mathbf v=\mathbf 0:\quad x+y=0\ \Rightarrow\ \mathbf p_1=\begin{pmatrix}1\\-1\end{pmatrix}\)

    One equation is enough: the two rows say the same thing. Any non-zero multiple of the vector works.

  2. Eigenvector for λ = 5

    \((\mathbf M-5\mathbf I)\mathbf v=\mathbf 0:\quad -3x+y=0\ \Rightarrow\ \mathbf p_2=\begin{pmatrix}1\\3\end{pmatrix}\)

    One equation is enough: the two rows say the same thing. Any non-zero multiple of the vector works.

  3. Check

    \(\mathbf M\mathbf p_2=\begin{pmatrix}2+3\\3+12\end{pmatrix}=5\begin{pmatrix}1\\3\end{pmatrix}\)

    Use the second row to check the eigenvector.

  4. General solution

    \(\begin{pmatrix}x\\y\end{pmatrix}=Ae^{t}\begin{pmatrix}1\\-1\end{pmatrix}+Be^{5t}\begin{pmatrix}1\\3\end{pmatrix}\)

    A and B are two arbitrary constants. Each eigenvalue goes in its own exponent.

Video walkthrough coming
Using x(0) and y(0)

The system \(x'=-x+y,\ y'=4x+2y\) has eigenvalues \(-2\) and \(3\). Find \(x(t)\) and \(y(t)\) given that \(x(0)=7\) and \(y(0)=3\).

Working
  1. Eigenvectors

    \(\lambda=-2:\ \mathbf p_1=\begin{pmatrix}1\\-1\end{pmatrix};\quad \lambda=3:\ \mathbf p_2=\begin{pmatrix}1\\4\end{pmatrix}\)

    For λ = 3: (−1 − 3)x + y = 0 gives y = 4x.

  2. General solution

    \(\begin{pmatrix}x\\y\end{pmatrix}=Ae^{-2t}\begin{pmatrix}1\\-1\end{pmatrix}+Be^{3t}\begin{pmatrix}1\\4\end{pmatrix}\)

    Write out both rows to get x and y separately.

  3. Put t = 0

    \(A+B=7,\quad -A+4B=3\)

    e⁰ = 1, so the exponentials disappear. One equation for x(0), one for y(0).

  4. Solve for A and B

    \(A=5,\ B=2\)

    Solve for A and B: add the equations to get 5B = 10.

  5. Particular solution

    \(x=5e^{-2t}+2e^{3t},\quad y=-5e^{-2t}+8e^{3t}\)

    Check: x(0) = 5 + 2 = 7 and y(0) = −5 + 8 = 3.

Video walkthrough coming
Using x(0) and x'(0)

The system \(x'=3x+2y,\ y'=x+2y\) has eigenvalues 1 and 4. Given that \(x(0)=5\) and \(x'(0)=17\), find \(x(t)\).

Working
  1. Find y(0)

    \(x'(0)=3x(0)+2y(0):\ 17=15+2y(0)\Rightarrow y(0)=1\)

    Put t = 0 in dx/dt = 3x + 2y. This turns the x' condition into a y(0) condition.

  2. Eigenvectors

    \(\lambda=1:\ \mathbf p_1=\begin{pmatrix}1\\-1\end{pmatrix};\quad\lambda=4:\ \mathbf p_2=\begin{pmatrix}2\\1\end{pmatrix}\)

    λ = 4: (3 − 4)x + 2y = 0 gives x = 2y.

  3. General solution

    \(\begin{pmatrix}x\\y\end{pmatrix}=Ae^{t}\begin{pmatrix}1\\-1\end{pmatrix}+Be^{4t}\begin{pmatrix}2\\1\end{pmatrix}\)

    x = A e^t + 2B e^{4t} and y = −A e^t + B e^{4t}.

  4. Put t = 0

    \(A+2B=5,\quad -A+B=1\ \Rightarrow\ A=1,\ B=2\)

    Solve for A and B (from the calculator, or add the equations).

  5. Answer

    \(x(t)=e^{t}+4e^{4t}\)

    Check: x'(0) = 1 + 16 = 17.

Video walkthrough coming