Stability Criteria

Stability criteria are the rules used to decide whether a linear system of differential equations settles back to equilibrium or moves away from it. In this course, they usually come from the eigenvalues of the system matrix.

What are Stability Criteria?

Stability criteria are the conditions you use in Linear Algebra and Differential Equations to decide what happens near an equilibrium point in a linear system. The big question is simple: if you nudge the system a little, does it come back, stay balanced, or drift farther away?

For a system written in matrix form, x' = Ax, the matrix A controls the motion. If every eigenvalue of A has a negative real part, the solution decays toward equilibrium as time goes on, so the equilibrium is asymptotically stable. If at least one eigenvalue has a positive real part, the exponential factor grows and the system becomes unstable.

That eigenvalue test is the quickest stability criterion for linear systems because eigenvalues summarize the long-term behavior of the matrix exponential e^{At}. A negative real part means the corresponding mode shrinks. A positive real part means it grows. Purely imaginary eigenvalues sit on the border, since they can produce oscillation without decay, so you may need more analysis before calling the equilibrium stable or unstable.

This is where the course gets a little more subtle than a basic definition. Some systems are stable but not asymptotically stable, which means solutions stay near equilibrium but do not necessarily move closer to it. In a linear setting, repeated zero or imaginary eigenvalues can make the conclusion depend on the matrix structure, not just the raw eigenvalue list.

You may also see stability criteria stated without explicitly finding eigenvalues. The Routh-Hurwitz criterion, for example, checks the signs and combinations of polynomial coefficients in the characteristic equation. That is useful when the eigenvalues are messy to compute by hand, but you still want to know whether the system is stable.

A good way to read these problems is to ask what each eigenmode does over time. Stable systems damp out, unstable systems amplify, and borderline cases need a second look because the algebra alone may not tell the whole story.

Why Stability Criteria matter in Linear Algebra and Differential Equations

Stability criteria give you the main interpretation step in linear differential equations: after you solve or set up the system, you still need to know what the solution means over time. A solution formula is not enough by itself if you cannot tell whether the motion approaches equilibrium, oscillates around it, or runs away.

This shows up constantly in the matrix method for systems, especially when you write a differential equation as x' = Ax and analyze A. The eigenvalues of A do more than help with diagonalization or computation. They tell you whether the model behaves like a damped system, a growing system, or a borderline case that needs extra checking.

That makes stability criteria useful for every kind of interpretation question in the course. If you are given a phase-plane sketch, a matrix, or a characteristic polynomial, stability is often the final conclusion you are expected to make. It is also a fast way to check whether a proposed equilibrium is realistic in a modeling problem, since unstable behavior can mean the model predicts runaway growth or collapse.

If you later work with nonlinear systems, the same vocabulary still matters, but the analysis becomes more local and sometimes more delicate. Learning the linear case well gives you the baseline for those comparisons.

How Stability Criteria connect across the course

Eigenvalues

Eigenvalues are the main tool behind most stability tests in linear systems. Their real parts tell you whether each solution mode decays, grows, or sits on the edge of stability. When you see a matrix differential equation, checking eigenvalues is often the first move before you make any claim about long-term behavior.

Equilibrium Point

Stability criteria are always judged relative to an equilibrium point. The question is not just whether a solution exists, but what happens if you start near that equilibrium. In a system x' = Ax, the equilibrium is usually the origin, and stability tells you whether nearby solutions return there.

Matrix Exponential

The matrix exponential e^{At} explains why eigenvalues control stability. It turns the system matrix into the time-evolution operator, so growth or decay comes from the exponential behavior of each eigenmode. If you understand e^{At}, stability is no longer just a rule, it becomes the outcome of the solution formula.

Lyapunov's Stability Theorem

Lyapunov's Stability Theorem gives a broader framework for deciding stability, especially beyond the simplest linear cases. In linear systems, eigenvalue criteria are often enough. In more advanced problems, Lyapunov methods let you prove stability using a function that measures how the system changes over time.

Are Stability Criteria on the Linear Algebra and Differential Equations exam?

A problem set or quiz question will usually give you a system matrix, a characteristic polynomial, or a set of eigenvalues and ask you to classify the equilibrium. Your job is to connect the sign of the real parts to the behavior of the solutions, then state whether the system is asymptotically stable, unstable, or borderline. If the eigenvalues are awkward, you may be expected to use a criterion like Routh-Hurwitz instead of solving them exactly. The main scoring move is not just computing numbers, but making the correct stability conclusion from them.

Stability Criteria vs Equilibrium Point

An equilibrium point is the state where the system does not change, while stability criteria tell you what happens after a small disturbance from that state. So the equilibrium is the object you study, and stability is the behavior you test around it.

Key things to remember about Stability Criteria

• Stability criteria tell you whether a linear differential system returns to equilibrium or moves away from it after a small disturbance.

• For a system x' = Ax, the eigenvalues of A usually give the fastest stability check.

• If all eigenvalues have negative real parts, the equilibrium is asymptotically stable.

• If any eigenvalue has a positive real part, the system is unstable.

• Purely imaginary or zero eigenvalues are borderline cases, so you may need more than the basic eigenvalue test.

Frequently asked questions about Stability Criteria

What is stability criteria in Linear Algebra and Differential Equations?

Stability criteria are the conditions used to decide whether a linear system of differential equations stays near an equilibrium or moves away from it. In the matrix form x' = Ax, this usually means checking the eigenvalues of A. The sign of the real part tells you whether each mode decays, grows, or stays borderline.

How do eigenvalues determine stability?

Eigenvalues tell you how each solution mode behaves over time. Negative real parts mean decay toward equilibrium, positive real parts mean growth away from equilibrium, and purely imaginary values usually mean oscillation or a borderline case. That is why eigenvalues are the standard shortcut for stability questions.

What is the difference between stable and asymptotically stable?

Stable means nearby solutions stay near the equilibrium, while asymptotically stable means they actually move toward it as time goes on. In linear systems, all eigenvalues having negative real parts gives asymptotic stability. If the system only avoids drifting away, but does not converge, it is stable but not asymptotically stable.

Can you test stability without finding eigenvalues?

Yes, sometimes. The Routh-Hurwitz criterion checks stability from the coefficients of the characteristic polynomial, which can be easier than solving for every eigenvalue. That is especially helpful when the polynomial is higher degree or the roots are messy.