Reference: Matrix Systems of Differential Equations - YouTube
Central idea: express the higher order DE into a first order DE with vectors.
General form:
a n x ( n ) + a n − 1 x ( n − 1 ) + ⋯ + a 2 x ¨ + a 1 x ˙ + a 0 x = 0 a_n x^{(n)} + a_{n-1} x^{(n-1)} + \dots + a_2 \ddot{x} + a_1 \dot{x} + a_0 x = 0 a n x ( n ) + a n − 1 x ( n − 1 ) + ⋯ + a 2 x ¨ + a 1 x ˙ + a 0 x = 0 Introduce new variables:
{ x 1 = x x 2 = x ˙ x 3 = x ¨ ⋮ x n = x ( n − 1 ) → take derivative x ˙ 1 = x ˙ = x 2 x ˙ 2 = x ¨ = x 3 x ˙ 3 = x ... = x 4 ⋮ x ˙ n = x ( n ) \begin{cases}
x_1 = x \\
x_2 = \dot{x} \\
x_3 = \ddot{x} \\
\vdots \\
x_n = x^{(n-1)}
\end{cases}
\quad \xrightarrow{\text{take derivative}} \quad
\begin{aligned}
\dot{x}_1 &= \dot{x} = x_2 \\
\dot{x}_2 &= \ddot{x} = x_3 \\
\dot{x}_3 &= \dddot{x} = x_4 \\
\vdots \\
\dot{x}_n &= x^{(n)}
\end{aligned} ⎩ ⎨ ⎧ x 1 = x x 2 = x ˙ x 3 = x ¨ ⋮ x n = x ( n − 1 ) take derivative x ˙ 1 x ˙ 2 x ˙ 3 ⋮ x ˙ n = x ˙ = x 2 = x ¨ = x 3 = x ... = x 4 = x ( n ) x ( n ) = − a 0 x − a 1 x ˙ − a 2 x ¨ − ⋯ − a n − 1 x ( n − 1 ) = − a 0 x 1 − a 1 x 2 − a 2 x 3 − ⋯ − a n − 1 x n \begin{aligned}
x^{(n)} &= -a_0 x - a_1 \dot{x} - a_2 \ddot{x} - \dots - a_{n-1} x^{(n-1)} \\
&= -a_0 x_1 - a_1 x_2 - a_2 x_3 - \dots - a_{n-1} x_n
\end{aligned} x ( n ) = − a 0 x − a 1 x ˙ − a 2 x ¨ − ⋯ − a n − 1 x ( n − 1 ) = − a 0 x 1 − a 1 x 2 − a 2 x 3 − ⋯ − a n − 1 x n d d t [ x 1 x 2 x 3 ⋮ x n − 1 x n ] = [ 0 1 0 … 0 0 0 0 1 … 0 0 0 0 0 … 0 0 ⋮ ⋮ ⋮ ⋱ ⋮ ⋮ 0 0 0 … 0 1 − a 0 − a 1 − a 2 … − a n − 2 − a n − 1 ] [ x 1 x 2 x 3 ⋮ x n − 1 x n ] \frac{d}{dt}
\begin{bmatrix}
x_1 \\
x_2 \\
x_3 \\
\vdots \\
x_{n-1} \\
x_n
\end{bmatrix}
=
\begin{bmatrix}
0 & 1 & 0 & \dots & 0 & 0 \\
0 & 0 & 1 & \dots & 0 & 0 \\
0 & 0 & 0 & \dots & 0 & 0 \\
\vdots & \vdots & \vdots & \ddots & \vdots & \vdots \\
0 & 0 & 0 & \dots & 0 & 1 \\
-a_0 & -a_1 & -a_2 & \dots & -a_{n-2} & -a_{n-1}
\end{bmatrix}
\begin{bmatrix}
x_1 \\
x_2 \\
x_3 \\
\vdots \\
x_{n-1} \\
x_n
\end{bmatrix} d t d x 1 x 2 x 3 ⋮ x n − 1 x n = 0 0 0 ⋮ 0 − a 0 1 0 0 ⋮ 0 − a 1 0 1 0 ⋮ 0 − a 2 … … … ⋱ … … 0 0 0 ⋮ 0 − a n − 2 0 0 0 ⋮ 1 − a n − 1 x 1 x 2 x 3 ⋮ x n − 1 x n Let x = [ x 1 x 2 ⋮ x n ] \mathbf{x} = \begin{bmatrix} x_1 \\ x_2 \\ \vdots \\ x_n \end{bmatrix} x = x 1 x 2 ⋮ x n and A = [ 0 1 0 … 0 0 0 0 1 … 0 0 0 0 0 … 0 0 ⋮ ⋮ ⋮ ⋱ ⋮ ⋮ 0 0 0 … 0 1 − a 0 − a 1 − a 2 … − a n − 2 − a n − 1 ] A =
\begin{bmatrix}
0 & 1 & 0 & \dots & 0 & 0 \\
0 & 0 & 1 & \dots & 0 & 0 \\
0 & 0 & 0 & \dots & 0 & 0 \\
\vdots & \vdots & \vdots & \ddots & \vdots & \vdots \\
0 & 0 & 0 & \dots & 0 & 1 \\
-a_0 & -a_1 & -a_2 & \dots & -a_{n-2} & -a_{n-1}
\end{bmatrix} A = 0 0 0 ⋮ 0 − a 0 1 0 0 ⋮ 0 − a 1 0 1 0 ⋮ 0 − a 2 … … … ⋱ … … 0 0 0 ⋮ 0 − a n − 2 0 0 0 ⋮ 1 − a n − 1 , then x ˙ = A x \dot{\mathbf{x}} = A\mathbf{x} x ˙ = A x , the eigenvalues of A A A are roots of characteristic polynomial.
Solving x ‾ ˙ = A ‾ ‾ x ‾ \dot{\underline{x}} = \underline{\underline{A}} \space \underline{x} x ˙ = A x with Eigenvalues and Eigenvectors# Example#
x r x_r x r stands for a population, and it grows follows the trend of x ˙ r = λ r x r \dot{x}_r = \lambda_r x_r x ˙ r = λ r x r .
Write these equations in a system:
{ x ˙ 1 = λ 1 x 1 x ˙ 2 = λ 2 x 2 ⋮ x ˙ n = λ n x n ⟹ d d t [ x 1 x 2 ⋮ x n ] = [ λ 1 0 … 0 0 λ 2 … 0 ⋮ ⋮ ⋱ ⋮ 0 0 … λ n ] ⏟ D ‾ ‾ [ x 1 x 2 ⋮ x n ] \begin{cases}
\dot{x}_1 = \lambda_1 x_1 \\
\dot{x}_2 = \lambda_2 x_2 \\
\vdots \\
\dot{x}_n = \lambda_n x_n
\end{cases}
\implies
\frac{d}{dt}
\begin{bmatrix}
x_1 \\
x_2 \\
\vdots \\
x_n
\end{bmatrix}
=
\underbrace{
\begin{bmatrix}
\lambda_1 & 0 & \dots & 0 \\
0 & \lambda_2 & \dots & 0 \\
\vdots & \vdots & \ddots & \vdots \\
0 & 0 & \dots & \lambda_n
\end{bmatrix}
}_{\underline{\underline{D}}}
\begin{bmatrix}
x_1 \\
x_2 \\
\vdots \\
x_n
\end{bmatrix} ⎩ ⎨ ⎧ x ˙ 1 = λ 1 x 1 x ˙ 2 = λ 2 x 2 ⋮ x ˙ n = λ n x n ⟹ d t d x 1 x 2 ⋮ x n = D λ 1 0 ⋮ 0 0 λ 2 ⋮ 0 … … ⋱ … 0 0 ⋮ λ n x 1 x 2 ⋮ x n Integrate (Solve x x x ):
{ x 1 ( t ) = x 1 ( 0 ) e λ 1 t x 2 ( t ) = x 2 ( 0 ) e λ 2 t ⋮ x n ( t ) = x n ( 0 ) e λ n t ⟹ [ x 1 x 2 ⋮ x n ] ( t ) = [ e λ 1 t 0 … 0 0 e λ 2 t … 0 ⋮ ⋮ ⋱ ⋮ 0 0 … e λ n t ] ⏟ e D ‾ ‾ t [ x 1 x 2 ⋮ x n ] ( 0 ) \begin{cases}
x_1(t) = x_1(0) e^{\lambda_1 t} \\
x_2(t) = x_2(0) e^{\lambda_2 t} \\
\vdots \\
x_n(t) = x_n(0) e^{\lambda_n t}
\end{cases}
\implies
\begin{bmatrix}
x_1 \\
x_2 \\
\vdots \\
x_n
\end{bmatrix}(t)
=
\underbrace{
\begin{bmatrix}
e^{\lambda_1 t} & 0 & \dots & 0 \\
0 & e^{\lambda_2 t} & \dots & 0 \\
\vdots & \vdots & \ddots & \vdots \\
0 & 0 & \dots & e^{\lambda_n t}
\end{bmatrix}
}_{e^{\underline{\underline{D}} t}}
\begin{bmatrix}
x_1 \\
x_2 \\
\vdots \\
x_n
\end{bmatrix}(0) ⎩ ⎨ ⎧ x 1 ( t ) = x 1 ( 0 ) e λ 1 t x 2 ( t ) = x 2 ( 0 ) e λ 2 t ⋮ x n ( t ) = x n ( 0 ) e λ n t ⟹ x 1 x 2 ⋮ x n ( t ) = e D t e λ 1 t 0 ⋮ 0 0 e λ 2 t ⋮ 0 … … ⋱ … 0 0 ⋮ e λ n t x 1 x 2 ⋮ x n ( 0 )
This is a very special (ideal) case, which the coefficient matrix e D ‾ t {e^{\underline{D} t}} e D t is a diagonal matrix.
For general system x ‾ ˙ = A ‾ ‾ x ‾ \dot{\underline{x}} = \underline{\underline{A}} \space \underline{x} x ˙ = A x #
Transform the coordinate: x ‾ = T ‾ ‾ z ‾ \underline{x} = \underline{\underline{T}} \space \underline{z} x = T z
Express x ‾ \underline{x} x in terms of z ‾ \underline{z} z
x ‾ ˙ = A ‾ ‾ x ‾ \dot{\underline{x}} = \underline{\underline{A}} \space \underline{x} x ˙ = A x
z ‾ ˙ = D ‾ ‾ z ‾ \dot{\underline{z}} = \underline{\underline{D}} \space \underline{z} z ˙ = D z
D ‾ ‾ \underline{\underline{D}} D is a n × n n \times n n × n diagonal matrix.
z ‾ ( t ) = e D ‾ ‾ t z ‾ ( 0 ) \underline{z}(t) = e^{\underline{\underline{D}} t} \underline{z}(0) z ( t ) = e D t z ( 0 )
x ‾ = T ‾ z ‾ x ‾ ˙ = T ‾ ‾ z ‾ ˙ T ‾ ‾ z ‾ ˙ = A ‾ ‾ T ‾ ‾ z ‾ z ‾ ˙ = T ‾ ‾ − 1 A ‾ ‾ T ‾ ‾ z ‾ \begin{aligned}
\underline{x} &= \underline{T} \space \underline{z} \\
\dot{\underline{x}} &= \underline{\underline{T}} \dot{\underline{z}} \\
\underline{\underline{T}} \dot{\underline{z}} &= \underline{\underline{A}} \underline{\underline{T}} \underline{z} \\
\dot{\underline{z}} &= \underline{\underline{T}}^{-1} \underline{\underline{A}} \underline{\underline{T}} \underline{z}
\end{aligned} x x ˙ T z ˙ z ˙ = T z = T z ˙ = A T z = T − 1 A T z
由于 z ‾ ˙ = D ‾ ‾ z ‾ \dot{\underline{z}} = \underline{\underline{D}} \underline{z} z ˙ = D z :找一个 T ‾ ‾ \underline{\underline{T}} T matrix,使 T ‾ ‾ − 1 A ‾ ‾ T ‾ ‾ \underline{\underline{T}}^{-1} \underline{\underline{A}} \underline{\underline{T}} T − 1 A T 满足 D ‾ ‾ \underline{\underline{D}} D .
We want T ‾ \underline{T} T so that:
T ‾ ‾ − 1 A ‾ ‾ T ‾ ‾ = D ‾ ‾ A ‾ = T ‾ D ‾ T ‾ − 1 \begin{aligned}
\underline{\underline{T}}^{-1} \underline{\underline{A}} \underline{\underline{T}} &= \underline{\underline{D}} \\
\underline{A} &= \underline{T} \underline{D} \underline{T}^{-1}
\end{aligned} T − 1 A T A = D = T D T − 1 This is the eigenvalue equation, where:
columns of T ‾ ‾ \underline{\underline{T}} T are the eigenvectors of A ‾ ‾ \underline{\underline{A}} A
diagonal elements of D ‾ ‾ \underline{\underline{D}} D are eigenvalues of A ‾ ‾ \underline{\underline{A}} A
[ A ‾ ‾ ] = [ ∣ ∣ ∣ t ‾ 1 t ‾ 2 … t ‾ n ∣ ∣ ∣ ] [ λ 1 0 λ 2 ⋱ 0 λ n ] [ ∣ ∣ ∣ t ‾ 1 t ‾ 2 … t ‾ n ∣ ∣ ∣ ] − 1 \begin{bmatrix}
& & & \\
& \underline{\underline{A}} & & \\
& & &
\end{bmatrix}
=
\begin{bmatrix}
| & | & & | \\
\underline{t}_1 & \underline{t}_2 & \dots & \underline{t}_n \\
| & | & & |
\end{bmatrix}
\begin{bmatrix}
\lambda_1 & & 0 \\
& \lambda_2 & \\
& & \ddots & \\
0 & & & \lambda_n
\end{bmatrix}
\begin{bmatrix}
| & | & & | \\
\underline{t}_1 & \underline{t}_2 & \dots & \underline{t}_n \\
| & | & & |
\end{bmatrix}^{-1} A = ∣ t 1 ∣ ∣ t 2 ∣ … ∣ t n ∣ λ 1 0 λ 2 0 ⋱ λ n ∣ t 1 ∣ ∣ t 2 ∣ … ∣ t n ∣ − 1 OR. (without the need to find inverse matrix).
[ A ‾ ‾ ] [ ∣ ∣ ∣ t ‾ 1 t ‾ 2 … t ‾ n ∣ ∣ ∣ ] = [ λ 1 0 λ 2 ⋱ 0 λ n ] [ ∣ ∣ ∣ t ‾ 1 t ‾ 2 … t ‾ n ∣ ∣ ∣ ] \begin{bmatrix}
& & & \\
& \underline{\underline{A}} & & \\
& & &
\end{bmatrix}
\begin{bmatrix}
| & | & & | \\
\underline{t}_1 & \underline{t}_2 & \dots & \underline{t}_n \\
| & | & & |
\end{bmatrix}
=
\begin{bmatrix}
\lambda_1 & & 0 \\
& \lambda_2 & \\
& & \ddots & \\
0 & & & \lambda_n
\end{bmatrix}
\begin{bmatrix}
| & | & & | \\
\underline{t}_1 & \underline{t}_2 & \dots & \underline{t}_n \\
| & | & & |
\end{bmatrix} A ∣ t 1 ∣ ∣ t 2 ∣ … ∣ t n ∣ = λ 1 0 λ 2 0 ⋱ λ n ∣ t 1 ∣ ∣ t 2 ∣ … ∣ t n ∣ = [ ∣ ∣ ∣ λ 1 t ‾ 1 λ 2 t ‾ 2 … λ n t ‾ n ∣ ∣ ∣ ] =
\begin{bmatrix}
| & | & & | \\
\lambda_1 \underline{t}_1 & \lambda_2 \underline{t}_2 & \dots & \lambda_n \underline{t}_n \\
| & | & & |
\end{bmatrix} = ∣ λ 1 t 1 ∣ ∣ λ 2 t 2 ∣ … ∣ λ n t n ∣ A ‾ ‾ t k = λ k t ‾ k \underline{\underline{A}} t_k = \lambda_k \underline{t}_k A t k = λ k t k .