Skip to content

22

约 1454 字大约 5 分钟

2025-02-28

problem

dxdt=Ax,A=(5113)\frac{d \vec{x}}{d t}=A \vec{x}, A=\begin{pmatrix}-5 & -1 \\ 1 & -3\end{pmatrix} ,求其通解,并写出 eAte^{A t}

solution

矩阵 AA 的特征方程为:

det(AλI)=5λ113λ=(λ+4)2=0\det(A - \lambda I) = \begin{vmatrix} -5-\lambda & -1 \\ 1 & -3-\lambda \end{vmatrix} = (\lambda + 4)^2 = 0

解得重特征值:

λ=4(二重根)\lambda = -4 \quad (\text{二重根})

解方程 (A+4I)v=0(A + 4I)\vec{v} = 0

A+4I=(1111)A + 4I = \begin{pmatrix} -1 & -1 \\ 1 & 1 \end{pmatrix}

可得线性方程 v1v2=0-v_1 - v_2 = 0,解为 v1=(11)\vec{v}_1 = \begin{pmatrix} 1 \\ -1 \end{pmatrix}

解方程 (A+4I)v=v1(A + 4I)\vec{v} = \vec{v}_1: 取 v2=(01)\vec{v}_2 = \begin{pmatrix} 0 \\ -1 \end{pmatrix}

通解形式为:

x(t)=e4t(C1v1+C2(tv1+v2))\vec{x}(t) = e^{-4t} \left(C_1 \vec{v}_1 + C_2 (t \vec{v}_1 + \vec{v}_2)\right)

具体展开为:

x(t)=e4t(C1(11)+C2(t(11)+(01)))\vec{x}(t) = e^{-4t} \left(C_1 \begin{pmatrix} 1 \\ -1 \end{pmatrix} + C_2 \left(t \begin{pmatrix} 1 \\ -1 \end{pmatrix} + \begin{pmatrix} 0 \\ -1 \end{pmatrix} \right)\right)

利用幂零矩阵分解:A=4I+NA = -4I + N,其中 N=(1111)N = \begin{pmatrix} -1 & -1 \\ 1 & 1 \end{pmatrix} 满足 N2=0N^2 = 0,则:

eAt=e4t(I+Nt)e^{A t} = e^{-4t}(I + N t)

具体展开为:

eAt=e4t(1ttt1+t)e^{A t} = e^{-4t} \begin{pmatrix} 1 - t & -t \\ t & 1 + t \end{pmatrix}

problem

dxdt=Ax,A=(1111)\frac{d \vec{x}}{d t}=A \vec{x}, A=\begin{pmatrix}1 & -1 \\ 1 & 1\end{pmatrix} ,求其通解,并写出 eAte^{A t}

solution

矩阵 AA 的特征方程为:

det(AλI)=0\det(A - \lambda I) = 0

代入矩阵 AA

det(1λ111λ)=(1λ)2+1\det\begin{pmatrix} 1 - \lambda & -1 \\ 1 & 1 - \lambda \end{pmatrix} = (1 - \lambda)^2 + 1

(1λ)2+1=0    (1λ)2=1    1λ=±i(1 - \lambda)^2 + 1 = 0 \implies (1 - \lambda)^2 = -1 \implies 1 - \lambda = \pm i

因此,特征值为:

λ1=1+i,λ2=1i\lambda_1 = 1 + i, \quad \lambda_2 = 1 - i

观察矩阵 AA 可分解为:

A=I+J,其中I=(1001), J=(0110)A = I + J, \quad \text{其中} \quad I = \begin{pmatrix} 1 & 0 \\ 0 & 1 \end{pmatrix}, \ J = \begin{pmatrix} 0 & -1 \\ 1 & 0 \end{pmatrix}

IIJJ 可交换,且 J2=IJ^2 = -I

利用 JJ 的幂级数展开:

eJt=costI+sintJ=(costsintsintcost)e^{J t} = \cos t \cdot I + \sin t \cdot J = \begin{pmatrix} \cos t & -\sin t \\ \sin t & \cos t \end{pmatrix}

由于 A=I+JA = I + JIIJJ 可交换:

eAt=eIteJt=et(costsintsintcost)e^{A t} = e^{I t} \cdot e^{J t} = e^t \cdot \begin{pmatrix} \cos t & -\sin t \\ \sin t & \cos t \end{pmatrix}

通解为:

x(t)=eAtC=et(costsintsintcost)(C1C2)\vec{x}(t) = e^{A t} \vec{C} = e^t \begin{pmatrix} \cos t & -\sin t \\ \sin t & \cos t \end{pmatrix} \begin{pmatrix} C_1 \\ C_2 \end{pmatrix}

分量为:

{x1(t)=et(C1costC2sint)x2(t)=et(C1sint+C2cost)\begin{cases} x_1(t) = e^t (C_1 \cos t - C_2 \sin t) \\ x_2(t) = e^t (C_1 \sin t + C_2 \cos t) \end{cases}

通解:

x(t)=et(C1costC2sintC1sint+C2cost),C1,C2R\vec{x}(t) = e^t \begin{pmatrix} C_1 \cos t - C_2 \sin t \\ C_1 \sin t + C_2 \cos t \end{pmatrix}, \quad C_1, C_2 \in \mathbb{R}

矩阵指数:

eAt=et(costsintsintcost)e^{A t} = e^t \begin{pmatrix} \cos t & -\sin t \\ \sin t & \cos t \end{pmatrix}

problem

dxdt=Ax,A=(121111101)\frac{d \vec{x}}{d t}=A \vec{x}, A=\begin{pmatrix}1 & -2 & -1 \\ -1 & 1 & 1 \\ 1 & 0 & -1\end{pmatrix} ,求其通解,并写出 eAte^{A t}

solution

矩阵 AA 的特征方程为:

det(AλI)=0\det(A - \lambda I) = 0

解得特征值:

λ1=0,λ2=2,λ3=1\lambda_1 = 0, \quad \lambda_2 = 2, \quad \lambda_3 = -1

解方程 (A0I)v=0(A - 0I)\vec{v} = 0,得特征向量:

v1=(101)\vec{v}_1 = \begin{pmatrix} 1 \\ 0 \\ 1 \end{pmatrix}

解方程 (A2I)v=0(A - 2I)\vec{v} = 0,得特征向量:

v2=(321)\vec{v}_2 = \begin{pmatrix} 3 \\ -2 \\ 1 \end{pmatrix}

解方程 (A+I)v=0(A + I)\vec{v} = 0,得特征向量:

v3=(012)\vec{v}_3 = \begin{pmatrix} 0 \\ 1 \\ -2 \end{pmatrix}

由于矩阵 AA 有三个不同的实特征值,通解为:

x(t)=C1e0tv1+C2e2tv2+C3etv3\vec{x}(t) = C_1 e^{0 \cdot t} \vec{v}_1 + C_2 e^{2t} \vec{v}_2 + C_3 e^{-t} \vec{v}_3

即:

x(t)=C1(101)+C2e2t(321)+C3et(012)\vec{x}(t) = C_1 \begin{pmatrix} 1 \\ 0 \\ 1 \end{pmatrix} + C_2 e^{2t} \begin{pmatrix} 3 \\ -2 \\ 1 \end{pmatrix} + C_3 e^{-t} \begin{pmatrix} 0 \\ 1 \\ -2 \end{pmatrix}

构造可逆矩阵 PP 和对角矩阵 DD

P=(130021112),D=(000020001)P = \begin{pmatrix} 1 & 3 & 0 \\ 0 & -2 & 1 \\ 1 & 1 & -2 \end{pmatrix}, \quad D = \begin{pmatrix} 0 & 0 & 0 \\ 0 & 2 & 0 \\ 0 & 0 & -1 \end{pmatrix}

矩阵指数为:

eAt=PeDtP1e^{A t} = P e^{D t} P^{-1}

其中:

eDt=(1000e2t000et)e^{D t} = \begin{pmatrix} 1 & 0 & 0 \\ 0 & e^{2t} & 0 \\ 0 & 0 & e^{-t} \end{pmatrix}

通过矩阵乘法计算 eAte^{A t},结果为:

eAt=16(3+3e2t66e2t33e2t2e2t+2et4e2t+2et2e2t2et3+e2t4et62e2t4et3e2t+4et)e^{A t} = \frac{1}{6} \begin{pmatrix} 3 + 3e^{2t} & 6 - 6e^{2t} & 3 - 3e^{2t} \\ -2e^{2t} + 2e^{-t} & 4e^{2t} + 2e^{-t} & 2e^{2t} - 2e^{-t} \\ 3 + e^{2t} - 4e^{-t} & 6 - 2e^{2t} - 4e^{-t} & 3 - e^{2t} + 4e^{-t} \end{pmatrix}

problem

dxdt=Ax,A=(110010104)\frac{d \vec{x}}{d t}=A \vec{x}, A=\begin{pmatrix}-1 & 1 & 0 \\ 0 & -1 & 0 \\ 1 & 0 & -4\end{pmatrix} ,求其通解,并写出 eAte^{A t}

solution

有特征值 λ1=4,λ2=λ3=1,\lambda_{1}=-4,\lambda_{2}=\lambda_{3}=-1, 特征向量 v=(0,0,1)\vec{v}=(0,0,1)^\prime

(Aλ2I)x=0(A-\lambda_{2}I)\vec{x}=0,有 v20=(301),v30=(130)\vec{v}_{20}=\begin{pmatrix}3 \\ 0 \\ 1\end{pmatrix}, \vec{v}_{30}=\begin{pmatrix}1 \\ 3 \\ 0\end{pmatrix}

v21=(Aλ2I)2v20=(000),v31=(Aλ2I)2v30=(301)\vec{v}_{21}=(A-\lambda_{2}I)^{2}\vec{v}_{20}=\begin{pmatrix}0 \\ 0 \\ 0\end{pmatrix}, \vec{v}_{31}=(A-\lambda_{2}I)^{2}\vec{v}_{30}=\begin{pmatrix}3 \\ 0 \\ 1\end{pmatrix}

故:

Φ(t)=(03et(1+3t)et003ete4tettet)\Phi(t)=\begin{pmatrix}0 & 3e^{-t} & (1+3t)e^{-t} \\ 0 & 0 & 3e^{-t} \\ e^{-4t} & e^{-t} & te^{-t}\end{pmatrix}

因此:

Φ1(0)=(13191131900130)\Phi^{-1}(0)=\begin{pmatrix} -\frac{1}{3} & \frac{1}{9} & 1 \\ \frac{1}{3} & -\frac{1}{9} & 0 \\ 0 & \frac{1}{3} & 0 \end{pmatrix}

矩阵指数:

eAt=Φ(t)Φ1(0)=(ettet00et0ete4t3tet3ete4t9e4t)e^{At}=\Phi(t)\Phi^{-1}(0)=\begin{pmatrix} e^{-t} & te^{-t} & 0 \\ 0 & e^{-t} & 0 \\ \frac{e^{-t} - e^{-4t}}{3} & \frac{te^{-t}}{3} - \frac{e^{-t} - e^{-4t}}{9} & e^{-4t} \end{pmatrix}

通解:

x(t)=C1et(301)+C2et(3t+13t)+C3e4t(001)\vec{x}(t) = C_1 e^{-t} \begin{pmatrix} 3 \\ 0 \\ 1 \end{pmatrix} + C_2 e^{-t} \begin{pmatrix} 3t + 1 \\ 3 \\ t \end{pmatrix} + C_3 e^{-4t} \begin{pmatrix} 0 \\ 0 \\ 1 \end{pmatrix}

problem

dxdt=Ax,A=(211212112)\frac{d \vec{x}}{d t}=A \vec{x}, A=\begin{pmatrix}2 & -1 & -1 \\ 2 & -1 & -2 \\ -1 & 1 & 2\end{pmatrix} ,求其通解,并写出 eAte^{A t}

solution

有特征值 λ1=λ2=λ3=1\lambda_{1}=\lambda_{2}=\lambda_{3}=1

(Aλ1)x=0(A-\lambda_{1})\vec{x}=0,有:

v10=(100),v20=(010),v30=(001)\vec{v}_{10}=\begin{pmatrix}1 \\ 0 \\ 0\end{pmatrix}, \vec{v}_{20}=\begin{pmatrix}0 \\ 1 \\ 0\end{pmatrix}, \vec{v}_{30}=\begin{pmatrix}0 \\ 0 \\ 1\end{pmatrix}

故:

v11=(Aλ1)v10=(121)\vec{v}_{11}=(A-\lambda_{1})\vec{v}_{10}=\begin{pmatrix}1 \\ 2 \\ -1\end{pmatrix}

v12=(Aλ1)2v10=(000)\vec{v}_{12}=(A-\lambda_{1})^{2}\vec{v}_{10}=\begin{pmatrix}0 \\ 0 \\ 0\end{pmatrix}

v21=(Aλ1)v20=(121)\vec{v}_{21}=(A-\lambda_{1})\vec{v}_{20}=\begin{pmatrix}-1 \\ -2 \\ 1\end{pmatrix}

v22=(Aλ1)2v20=(000)\vec{v}_{22}=(A-\lambda_{1})^{2}\vec{v}_{20}=\begin{pmatrix}0 \\ 0 \\ 0\end{pmatrix}

v31=(Aλ1)v30=(121)\vec{v}_{31}=(A-\lambda_{1})\vec{v}_{30}=\begin{pmatrix}-1 \\ -2 \\ 1\end{pmatrix}

v32=(Aλ1)2v30=(000)\vec{v}_{32}=(A-\lambda_{1})^{2}\vec{v}_{30}=\begin{pmatrix}0 \\ 0 \\ 0\end{pmatrix}

则:

Φ(t)=et(1+ttt2t12t2ttt1+t)\Phi(t)=e^t\begin{pmatrix} 1 + t & -t & -t \\ 2t & 1 - 2t & -2t \\ -t & t & 1 + t \end{pmatrix}

矩阵指数:

eAt=Φ(t)Φ1(0)=et(1+ttt2t12t2ttt1+t)e^{A t} =\Phi(t)\Phi^{-1}(0)= e^t \begin{pmatrix} 1 + t & -t & -t \\ 2t & 1 - 2t & -2t \\ -t & t & 1 + t \end{pmatrix}

通解:

x(t)=eAtc\vec{x}(t) = e^{At}\vec{c}

贡献者: Ming Wei