4.3 Phase Portraits and Stability Analysis

Understanding the qualitative behavior of 2D linear systems through geometric visualization

Learning Objectives

Understand phase portraits as geometric representations of system trajectories
Classify equilibrium points based on eigenvalue analysis
Determine stability of the origin as an equilibrium point
Use the trace-determinant plane to classify linear systems
Interpret and sketch phase portraits for 2×2 systems

What is a Phase Portrait?

A phase portrait is a plot of trajectories in the $x_1$-$x_2$ phase plane for a dynamical system. For a 2×2 linear system

$$\mathbf{x}' = A\mathbf{x}$$

each solution $\mathbf{x}(t) = [x_1(t), x_2(t)]^T$ traces a curve (trajectory) in the phase plane as $t$ varies. The arrows on these curves indicate the direction of motion as time increases.

A phase portrait captures the qualitative behavior of the system:

What is Stability?

An equilibrium point is a point where $\mathbf{x}' = \mathbf{0}$. For the linear system $\mathbf{x}' = A\mathbf{x}$, the origin $\mathbf{x} = \mathbf{0}$ is always an equilibrium.

Stability Definitions

The origin is:

  • Stable if for any $\epsilon > 0$, there exists $\delta > 0$ such that $\|\mathbf{x}_0\| < \delta$ implies $\|\mathbf{x}(t)\| < \epsilon$ for all $t \geq 0$.
  • Asymptotically stable if it is stable AND $\lim_{t \to \infty} \mathbf{x}(t) = \mathbf{0}$ for all nearby initial conditions.
  • Unstable if it is not stable.

Key insight: For 2×2linear systems, the stability of the origin is completely determined by the eigenvalues of $A$.

Classification of Equilibrium Points

The behavior of a 2×2 linear system is determined entirely by the eigenvalues $\lambda_1, \lambda_2$ of the coefficient matrix $A$. Below is a comprehensive classification:

Stable Node

$\lambda_1, \lambda_2 \in \mathbb{R}$, $\lambda_1 < \lambda_2 < 0$
All trajectories approach the origin along straight line directions (eigenvectors). Motion is exponential decay with rate $\max(|\lambda_1|, |\lambda_2|)$.
Asymptotically Stable

Unstable Node

$\lambda_1, \lambda_2 \in \mathbb{R}$, $0 < \lambda_1 < \lambda_2$
All trajectories diverge from the origin along eigenvector directions. Exponential growth away from origin.
Unstable

Saddle Point

$\lambda_1, \lambda_2 \in \mathbb{R}$, $\lambda_1 < 0 < \lambda_2$
The stable manifold (along $\mathbf{v}_1$) attracts trajectories; the unstable manifold (along $\mathbf{v}_2$) repels them. Most trajectories diverge.
Unstable

Stable Spiral (Sink)

$\lambda = \alpha \pm \beta i$, $\alpha < 0, \beta \neq 0$
Trajectories spiral inward toward the origin. The rotation rate is $\beta$ (rad/s), the decay rate is $e^{\alpha t}$.
Asymptotically Stable

Unstable Spiral (Source)

$\lambda = \alpha \pm \beta i$, $\alpha > 0, \beta \neq 0$
Trajectories spiral outward from the origin. Exponential growth with oscillatory behavior.
Unstable

Center (Elliptic)

$\lambda = \pm \beta i$, $\beta > 0$ (purely imaginary)
All trajectories form closed ellipses around the origin. The system oscillates with constant period $T = 2\pi/\beta$.
Stable (but NOT asymptotically)

Star Node (Degenerate)

Repeated eigenvalue: $\lambda$ with 2 independent eigenvectors ($A = \lambda I$)
All trajectories are straight lines through the origin. Either all converge (if $\lambda < 0$) or diverge (if $\lambda > 0$).
Stable if $\lambda < 0$

Improper Node

Repeated eigenvalue: $\lambda$ with only 1 independent eigenvector
Trajectories approach the origin along the eigenvector direction, but not straight lines. One "fast" and one "slow" mode.
Asymptotically Stable if $\lambda < 0$

Summary Table

Type Eigenvalues Condition Stability
Stable Node $\lambda_1, \lambda_2 \in \mathbb{R}$ $\lambda_1 < \lambda_2 < 0$ Asymptotically Stable
Unstable Node $\lambda_1, \lambda_2 \in \mathbb{R}$ $0 < \lambda_1 < \lambda_2$ Unstable
Saddle Point $\lambda_1, \lambda_2 \in \mathbb{R}$ $\lambda_1 < 0 < \lambda_2$ Unstable
Stable Spiral $\lambda = \alpha \pm \beta i$ $\alpha < 0, \beta \neq 0$ Asymptotically Stable
Unstable Spiral $\lambda = \alpha \pm \beta i$ $\alpha > 0, \beta \neq 0$ Unstable
Center $\lambda = \pm \beta i$ $\beta > 0$ Stable
Star Node Repeated $\lambda$ 2 independent eigenvectors Stable if $\lambda < 0$
Improper Node Repeated $\lambda$ 1 independent eigenvector Stable if $\lambda < 0$

The Trace-Determinant Plane

For a 2×2 matrix $A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}$, define:

$$\tau = \text{tr}(A) = a + d \quad \text{and} \quad \Delta = \det(A) = ad - bc$$

The characteristic polynomial is $\lambda^2 - \tau\lambda + \Delta = 0$, with discriminant $D = \tau^2 - 4\Delta$.

The $(\tau, \Delta)$ plane is divided into regions corresponding to different equilibrium types:

How to use the diagram: Compute $\tau$ and $\Delta$ for your system, plot the point $(\tau, \Delta)$. The region containing that point determines the equilibrium type.

Worked Examples

Solution

Step 1: Find eigenvalues

$$\det(A - \lambda I) = \det\begin{pmatrix} -3-\lambda & 1 \\ 0 & -2-\lambda \end{pmatrix} = (-3-\lambda)(-2-\lambda) = \lambda^2 + 5\lambda + 6 = 0$$

Factoring: $(\lambda + 2)(\lambda + 3) = 0$, so $\lambda_1 = -3, \lambda_2 = -2$.

Step 2: Check trace and determinant

$$\tau = -3 + (-2) = -5, \quad \Delta = (-3)(-2) - (1)(0) = 6$$

Point $(-5, 6)$ in the $(\tau, \Delta)$ plane: $\tau < 0$ and $\Delta > 0$ with $\tau^2 - 4\Delta = 25 - 24 = 1 > 0$ (real eigenvalues) → Stable Node

Step 3: Find eigenvectors

For $\lambda_1 = -3$: $(A + 3I)\mathbf{v} = \begin{pmatrix} 0 & 1 \\ 0 & 1 \end{pmatrix}\mathbf{v} = \mathbf{0}$ gives $\mathbf{v}_1 = \begin{pmatrix} 1 \\ 0 \end{pmatrix}$

For $\lambda_2 = -2$: $(A + 2I)\mathbf{v} = \begin{pmatrix} -1 & 1 \\ 0 & 0 \end{pmatrix}\mathbf{v} = \mathbf{0}$ gives $\mathbf{v}_2 = \begin{pmatrix} 1 \\ 1 \end{pmatrix}$

Step 4: Sketch phase portrait

Interpretation: Both eigenvalues are negative, so all trajectories decay exponentially to the origin. The decay is faster along the $v_1$ direction (rate $e^{-3t}$) than along the $v_2$ direction (rate $e^{-2t}$). The origin is asymptotically stable.

Solution

Step 1: Find eigenvalues

$$\det(A - \lambda I) = \det\begin{pmatrix} -\lambda & 1 \\ -1 & -\lambda \end{pmatrix} = \lambda^2 + 1 = 0$$

So $\lambda = \pm i$, purely imaginary with $\beta = 1$.

Step 2: Check trace and determinant

$$\tau = 0 + 0 = 0, \quad \Delta = (0)(0) - (1)(-1) = 1$$

Point $(0, 1)$ on the boundary $\tau = 0$, $\Delta > 0$ → Center

Step 3: Solution structure

$$\mathbf{x}(t) = c_1 \begin{pmatrix} \cos t \\ -\sin t \end{pmatrix} + c_2 \begin{pmatrix} \sin t \\ \cos t \end{pmatrix}$$

Step 4: Phase portrait

Interpretation: All trajectories are circles centered at the origin. The period is $T = 2\pi/1 = 2\pi$. The origin is stable but not asymptotically stable — nearby solutions stay near the origin but don't converge to it.

Solution

Step 1: Find eigenvalues

$$\det(A - \lambda I) = \det\begin{pmatrix} -1-\lambda & -4 \\ 1 & -1-\lambda \end{pmatrix} = (-1-\lambda)^2 + 4 = \lambda^2 + 2\lambda + 5 = 0$$
$$\lambda = \frac{-2 \pm \sqrt{4-20}}{2} = \frac{-2 \pm \sqrt{-16}}{2} = -1 \pm 2i$$

So $\alpha = -1 < 0$ and $\beta = 2 \neq 0$.

Step 2: Check trace and determinant

$$\tau = -1 + (-1) = -2, \quad \Delta = (-1)(-1) - (-4)(1) = 1 + 4 = 5$$

$\tau < 0$, $\Delta > 0$, $\tau^2 - 4\Delta = 4 - 20 = -16 < 0$ (complex eigenvalues with negative real part) → Stable Spiral

Step 3: Phase portrait

Interpretation: Trajectories spiral inward toward the origin. The decay rate is $e^{-t}$ and the rotation rate is $2$ rad/s. The origin is asymptotically stable. This is the underdamped case in oscillator problems.

Interactive Phase Portrait Explorer

Enter the elements of a 2×2 matrix $A$ to visualize its phase portrait. The system will automatically compute eigenvalues and classify the equilibrium type.

Trace (τ)
Determinant (Δ)
Eigenvalue 1
Eigenvalue 2
Classification

Exam-Style Practice Problems

Problem 1

Classify the equilibrium of the system $\mathbf{x}' = \begin{pmatrix} 1 & -3 \\ 3 & 1 \end{pmatrix}\mathbf{x}$ and describe the phase portrait behavior.

Problem 2

A mass-spring-damper system is described by $y'' + 4y' + 5y = 0$. Convert this to a first-order system, find its eigenvalues, and classify the equilibrium.

Problem 3 — True/False with Explanation

Statement: "A system with $\text{tr}(A) = 0$ and $\det(A) > 0$ must have a center at the origin."