Problem 74
Hard Phase-Plane Analysis

Solve the system of differential equations: dot x=x-y+z, dot…

Solve the system of differential equations: $$ \dot x=x-y+z,\qquad \dot y=x+y-z,\qquad \dot z=-y+2z $$

Explanation of Method

The eigenvalue $\lambda=1$ has algebraic multiplicity 2 but only a one-dimensional eigenspace, so a genuine Jordan block appears: a generalized eigenvector $\mathbf w$ solving $(A-I)\mathbf w=\mathbf v$ is required to complete a three-dimensional solution basis, producing a $te^t$ secular term.

  1. Write the system in matrix form and compute the characteristic polynomial.
  2. Identify the simple eigenvalue $\lambda=2$ and the repeated eigenvalue $\lambda=1$.
  3. Find the eigenvector at $\lambda=1$ and confirm the geometric multiplicity is only 1.
  4. Solve for a generalized eigenvector and build the Jordan-chain solution.
  5. Assemble and verify the general solution.

Step 1 · Matrix form and characteristic equation

With $A=\begin{pmatrix}1&-1&1\\1&1&-1\\0&-1&2\end{pmatrix}$, expanding $\det(A-\lambda I)$ gives

$$ -(\lambda-2)(\lambda-1)^2=0. $$

Eigenvalues: $\lambda_1=1$ (algebraic multiplicity 2) and $\lambda_2=2$ (simple).

Step 2 · Eigenvector for $\lambda_2=2$

$$ \mathbf v_1=\begin{pmatrix}1\\0\\1\end{pmatrix},\qquad A\mathbf v_1=2\mathbf v_1.\ \checkmark $$

Step 3 · Eigenvector for $\lambda_1=1$ and its rank deficiency

$A-I=\begin{pmatrix}0&-1&1\\1&0&-1\\0&-1&1\end{pmatrix}$ has rank 2 (rows 1 and 3 are identical), so the null space is only one-dimensional:

$$ \mathbf v_2=\begin{pmatrix}1\\1\\1\end{pmatrix},\qquad A\mathbf v_2=\mathbf v_2.\ \checkmark $$

Geometric multiplicity $1<2$: a generalized eigenvector is required.

Step 4 · Generalized eigenvector and Jordan-chain solutions

Solving $(A-I)\mathbf w=\mathbf v_2$ (consistent since $\mathbf v_2$ lies in the range of $A-I$) gives $\mathbf w=(1,-1,0)$, verified directly. The Jordan chain produces three independent solutions:

$$ \mathbf u^{(1)}=e^{2t}\mathbf v_1,\qquad \mathbf u^{(2)}=e^{t}\mathbf v_2,\qquad \mathbf u^{(3)}=e^{t}(t\,\mathbf v_2+\mathbf w). $$

Step 5 · General solution and verification

The three solutions are linearly independent at $t=0$ ($\det=-1\neq0$); direct substitution confirms the general solution below satisfies all three original equations exactly. ✓


Summary Table

Quantity Value
Eigenvalues$\lambda=1$ (mult.\ 2, geometric mult.\ 1), $\lambda=2$ (simple)
Generalized eigenvector$\mathbf w=(1,-1,0)$, solving $(A-I)\mathbf w=\mathbf v_2$
Structureone genuine $2\times2$ Jordan block at $\lambda=1$

Additional Notes

Result

The general solution of the system is:

$$ \begin{aligned} x(t) &= C_1\,e^{2t} + C_2\,e^{t} + C_3(t+1)e^{t},\\ y(t) &= C_2\,e^{t} + C_3(t-1)e^{t},\\ z(t) &= C_1\,e^{2t} + C_2\,e^{t} + C_3\,t\,e^{t}. \end{aligned} $$