841 字
4 分钟
Reducing Higher Order Linear ODE Systems
2026-10-29
无标签

Convert General Linear ODE into Matrix Form#

Reference: Matrix Systems of Differential Equations - YouTube

Central idea: express the higher order DE into a first order DE with vectors.

General form:

anx(n)+an−1x(n−1)+⋯+a2x¨+a1x˙+a0x=0a_n x^{(n)} + a_{n-1} x^{(n-1)} + \dots + a_2 \ddot{x} + a_1 \dot{x} + a_0 x = 0

Introduce new variables:

{x1=xx2=x˙x3=x¨⋮xn=x(n−1)→take derivativex˙1=x˙=x2x˙2=x¨=x3x˙3=x...=x4⋮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(n)=−a0x−a1x˙−a2x¨−⋯−an−1x(n−1)=−a0x1−a1x2−a2x3−⋯−an−1xn\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}ddt[x1x2x3⋮xn−1xn]=[010…00001…00000…00⋮⋮⋮⋱⋮⋮000…01−a0−a1−a2…−an−2−an−1][x1x2x3⋮xn−1xn]\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}

Let x=[x1x2⋮xn]\mathbf{x} = \begin{bmatrix} x_1 \\ x_2 \\ \vdots \\ x_n \end{bmatrix} and A=[010…00001…00000…00⋮⋮⋮⋱⋮⋮000…01−a0−a1−a2…−an−2−an−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}, then x˙=Ax\dot{\mathbf{x}} = A\mathbf{x}, the eigenvalues of AA are roots of characteristic polynomial.

Solving x‾˙=A‾‾ x‾\dot{\underline{x}} = \underline{\underline{A}} \space \underline{x} with Eigenvalues and Eigenvectors#

Example#

xrx_r stands for a population, and it grows follows the trend of x˙r=λrxr\dot{x}_r = \lambda_r x_r.

Write these equations in a system:

{x˙1=λ1x1x˙2=λ2x2⋮x˙n=λnxn  ⟹  ddt[x1x2⋮xn]=[λ10…00λ2…0⋮⋮⋱⋮00…λn]⏟D‾‾[x1x2⋮xn]\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}

Integrate (Solve xx):

{x1(t)=x1(0)eλ1tx2(t)=x2(0)eλ2t⋮xn(t)=xn(0)eλnt  ⟹  [x1x2⋮xn](t)=[eλ1t0…00eλ2t…0⋮⋮⋱⋮00…eλnt]⏟eD‾‾t[x1x2⋮xn](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)

This is a very special (ideal) case, which the coefficient matrix eD‾t{e^{\underline{D} t}} is a diagonal matrix.

For general system x‾˙=A‾‾ x‾\dot{\underline{x}} = \underline{\underline{A}} \space \underline{x}#

Transform the coordinate: x‾=T‾‾ z‾\underline{x} = \underline{\underline{T}} \space \underline{z}

Express x‾\underline{x} in terms of z‾\underline{z}

  1. x‾˙=A‾‾ x‾\dot{\underline{x}} = \underline{\underline{A}} \space \underline{x} z‾˙=D‾‾ z‾\dot{\underline{z}} = \underline{\underline{D}} \space \underline{z}

    D‾‾\underline{\underline{D}} is a n×nn \times n diagonal matrix.

    z‾(t)=eD‾‾tz‾(0)\underline{z}(t) = e^{\underline{\underline{D}} t} \underline{z}(0)

  2. x‾=T‾ z‾x‾˙=T‾‾z‾˙T‾‾z‾˙=A‾‾T‾‾z‾z‾˙=T‾‾−1A‾‾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}

由于 z‾˙=D‾‾z‾\dot{\underline{z}} = \underline{\underline{D}} \underline{z}:找一个 T‾‾\underline{\underline{T}} matrix,使 T‾‾−1A‾‾T‾‾\underline{\underline{T}}^{-1} \underline{\underline{A}} \underline{\underline{T}} 满足 D‾‾\underline{\underline{D}}. We want T‾\underline{T} so that:

T‾‾−1A‾‾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}

This is the eigenvalue equation, where:

  • columns of T‾‾\underline{\underline{T}} are the eigenvectors of A‾‾\underline{\underline{A}}
  • diagonal elements of D‾‾\underline{\underline{D}} are eigenvalues of A‾‾\underline{\underline{A}}
[A‾‾]=[∣∣∣t‾1t‾2…t‾n∣∣∣][λ10λ2⋱0λn][∣∣∣t‾1t‾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}

OR. (without the need to find inverse matrix).

[A‾‾][∣∣∣t‾1t‾2…t‾n∣∣∣]=[λ10λ2⋱0λn][∣∣∣t‾1t‾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}=[∣∣∣λ1t‾1λ2t‾2…λnt‾n∣∣∣]= \begin{bmatrix} | & | & & | \\ \lambda_1 \underline{t}_1 & \lambda_2 \underline{t}_2 & \dots & \lambda_n \underline{t}_n \\ | & | & & | \end{bmatrix}

A‾‾tk=λkt‾k\underline{\underline{A}} t_k = \lambda_k \underline{t}_k.

Reducing Higher Order Linear ODE Systems
https://thyrius.top/posts/math/reducing-high-order-ode/
作者
Thyrius
发布于
2026-10-29
许可协议
CC BY-NC-SA 4.0