Amplification factor
yₙ₊₁ = (1 − hλ) yₙ
Stable if |1 − hλ| < 1.
For the test equation (), Euler's method gives , which decays only if , i.e. .
This page is easier if you are comfortable with:
With the limit is . For the amplification factor is : the numerical solution drops to zero at once. For it is , so grows by a factor 1.5 per step with alternating sign, although the true solution decays. For the limit is : the faster the true decay, the smaller the step must be. The step formula is that of Euler's method.
This is the practical meaning of a stiff problem: a rapidly decaying component forces tiny steps even when we only care about slow behaviour. Implicit methods have no such limit, as in the Crank–Nicolson scheme; for the heat equation the same limit appears as (explicit heat scheme stability) and for waves as the CFL condition.
Higher-order methods have larger stability regions (classical Runge–Kutta, the midpoint method) but the same principle. The exponential decay model being tested is exponential growth and decay.
Euler's method only stays stable if the step is small compared with the fastest decay rate.
Above h = 0.2 the factor exceeds 1 and the numerical solution grows.
These examples are worked by hand, step by step, so you can compare your own working.
Find the largest stable step for .
Find the amplification factor for , .
Find the largest stable step for .
A few situations come up again and again. Knowing them saves time and helps you spot answers that cannot be right.
yₙ₊₁ = (1 − hλ) yₙ
Stable if |1 − hλ| < 1.
h < 2/λ; for λ = 10, h < 0.2
Boundary at h = 0.2.
h = 0.3, λ = 10: factor −2
|−2| > 1: grows.
e^(−hλ) = e^(−1) = 0.3679 for h = 0.1
Euler gives 0 here.
Applying forward Euler to y′ = −λy gives yₙ₊₁ = (1 − hλ)yₙ. The numerical solution decays only if |1 − hλ| < 1, i.e. h < 2/λ.
λ = 10, h = 0.1. Find the amplification factor.
Find the largest stable h for λ = 10.
h = 0.3, λ = 10. Find the factor.
Starting from y₀ = 1, find y₃ with h = 0.3, λ = 10.
stable if h < 1/λ ✗
stable if h < 2/λ ✓
The condition is |1 − hλ| < 1.
factor −2 decays ✗
|factor| must be < 1 ✓
Oscillating growth.
e^(−hλ) is the Euler factor ✗
it is the exact factor ✓
Euler factor is 1 − hλ.
any h works ✗
stiff problems need small h ✓
Use implicit methods.
Try each one on paper, then open the answer. Use the solver above to see the full working.
0.1
0
-1.5
2.25
0
0.02
Unlimited fresh questions with instant marking. Choose a level, work it out on paper, then type your answer.
Solve: …
When h < 2/λ.
1 − hλ.
The solution oscillates and grows.
One with a large λ forcing a tiny h.
classical Runge–Kutta.
explicit heat scheme stability.
More in Numerical methods and optimisation and the formula sheets.