Lax equivalence theorem

The lax equivalence theorem says a multistep method for an ODE converges when it is consistent and stable in the right sense. In this course, it’s the bridge between getting the formula right and getting a reliable approximation.

What is the lax equivalence theorem?

The lax equivalence theorem is the rule that connects three things you keep seeing in numerical differential equations: consistency, stability, and convergence. In Linear Algebra and Differential Equations, it tells you when a multistep method does more than just look reasonable on paper. If the method is consistent and stable enough, then its approximations actually move toward the true solution as the step size gets smaller.

The basic idea is simple. Consistency checks the formula itself. It asks whether, if you shrink the step size, the method matches the differential equation more and more closely. That is about local behavior, one step at a time. Stability is about what happens when small errors show up, which they always do because of rounding, starting values, or earlier approximation error.

This is where the theorem becomes useful. A method can be consistent and still fail badly if errors get amplified as you march forward in time. For multistep methods, the old error is reused in the next step, so instability can spread quickly. The lax equivalence theorem says that convergence is not just about the formula being a good approximation locally. You also need the method to control error propagation.

In the version used for this course topic, you often see the stability condition phrased in terms of the method’s stability region. For problems where the true solution should decay, the left half of the complex plane matters because eigenvalues with negative real part usually correspond to decaying modes. If the method’s stability region covers that part of the plane, it is much more likely to behave well on stiff or strongly damped systems.

A compact example is a linear multistep method applied to a test equation like y' = λy with Re(λ) < 0. If the method is consistent but its stability region misses that left-half-plane behavior, the computed solution can grow even though the true solution shrinks. That is exactly the kind of mismatch the theorem warns you about. So the theorem is not just a label, it is the checkpoint that tells you whether a multistep method is mathematically trustworthy for long runs.

Why the lax equivalence theorem matters in Linear Algebra and Differential Equations

This theorem matters because it gives you a clean way to judge a numerical method before you trust its output. In Differential Equations, you are often approximating solutions that cannot be written in closed form, or that are too expensive to solve exactly. The lax equivalence theorem tells you which properties actually need to be checked so you do not confuse a nice-looking formula with a reliable one.

It also connects directly to how multistep methods are built. These methods save work by reusing past function values, which makes them efficient, but that same feature makes them sensitive to accumulated error. The theorem explains why one bad stability choice can ruin an otherwise accurate method. That is a big deal in engineering and physics problems where you may compute many steps in a row.

The theorem also gives you vocabulary for talking about error. Consistency points to local truncation error, while convergence is about global error over many steps. Stability is the link that keeps local mistakes from spreading too far. If you understand the theorem, you can read a method description and immediately ask the right questions: Is it consistent? Is it stable? Does it converge in the situations I care about?

This is especially useful when a class compares different numerical schemes. You are not just memorizing formulas, you are comparing how they behave over time. The theorem is the reason a method can be perfectly valid algebraically but still be a poor choice for a stiff problem or a long interval of integration.

How the lax equivalence theorem connects across the course

Multistep Methods

The lax equivalence theorem is stated in the setting of multistep methods, where each new value depends on several earlier approximations. That reuse of past steps is why efficiency goes up, but it is also why error can spread from one step to the next. The theorem helps you decide whether that efficiency comes with reliable long-term behavior.

Consistency

Consistency checks whether the numerical formula matches the differential equation as the step size shrinks. On its own, that only tells you the method is locally accurate. The theorem says consistency is necessary, but not enough by itself, because the method still needs stability to produce convergence.

Stability

Stability is the part of the story that controls how errors behave after they appear. In multistep methods, a tiny error from rounding or a starting approximation can get reused many times. The theorem uses stability as the missing link between a correct local formula and a solution that stays close to the true one.

Convergence

Convergence is what you want at the end of the process: the numerical solution gets closer to the exact solution as the step size goes to zero. The lax equivalence theorem gives a practical criterion for convergence instead of forcing you to check everything from scratch. If consistency and the right stability condition hold, convergence follows.

Is the lax equivalence theorem on the Linear Algebra and Differential Equations exam?

A problem set question will usually give you a multistep method and ask you to decide whether it converges, or to explain why a stable-looking approximation still fails. You use the theorem by checking consistency first, then checking the stability condition or the stability region. If the method is consistent but not stable, you explain that local accuracy does not guarantee the global error stays under control. If the problem involves a test equation or a stiff system, you may also describe how the left half-plane enters the stability analysis. The move is less about memorizing a phrase and more about justifying whether the method is reliable over many steps.

The lax equivalence theorem vs Consistency

Consistency is only one piece of the story, while the lax equivalence theorem is the rule that connects consistency with stability and convergence. A method can be consistent and still produce bad long-term results if it is unstable. So if a question asks about the theorem, do not stop at the local error check.

Key things to remember about the lax equivalence theorem

• The lax equivalence theorem connects consistency, stability, and convergence for numerical methods for differential equations.

• In this course, it is most useful for multistep methods, where past approximations are reused in later steps.

• A method can be consistent without being reliable, because instability can make errors grow as you compute farther in time.

• The theorem gives you a fast way to judge whether a numerical method is likely to track the true solution well.

• For decaying solutions and stiff problems, the stability region and the left half-plane matter a lot.

Frequently asked questions about the lax equivalence theorem

What is the lax equivalence theorem in Linear Algebra and Differential Equations?

It is the statement that a multistep numerical method converges when it has the right consistency and stability properties. In this course, it is the bridge between a formula that approximates an ODE locally and a solution that stays accurate over many steps.

Is consistency enough for convergence?

No. Consistency only says the method matches the differential equation better as the step size gets smaller. You also need stability, because without it small errors can build up and stop the numerical solution from approaching the true one.

How does stability show up in multistep methods?

Stability shows up in how earlier errors affect later values. Since multistep methods reuse old approximations, any error that gets introduced can keep feeding into future steps. If the method’s stability region is poor, those errors can grow instead of fading away.

Why does the left half of the complex plane matter?

For many differential equations, especially ones with decaying behavior, eigenvalues with negative real part represent modes that should shrink over time. If a method is stable on the left half-plane, it is more likely to handle those problems without producing artificial growth.