Differential Equations · Undergraduate
Convolution and the Convolution Theorem
Quick answer
The Laplace transform of a product is not the product of transforms, but the transform of a convolution is. The convolution (f * g)(t) adds up f at each earlier time τ, weighted by g at the elapsed time t − τ. The convolution theorem L{f * g} = F(s)G(s) lets you invert a product such as 1/(s²(s + 1)) without partial fractions, and it gives the solution of y″ + y = g(t) for any forcing g as y = ∫₀ᵗ sin(t − τ) g(τ) dτ: the response to each instant of forcing, added up.
What you'll learn
- Compute convolutions of simple functions
- Use the convolution theorem to invert products of transforms
- Write the solution of a forced IVP as a convolution integral
- Solve simple integral equations with Laplace transforms
Products of transforms
Inverting is a common last step when solving IVPs, and the inverse of a product is not the product of inverses. It is the convolution:
Convolution is commutative, , and a few small cases are worth knowing: , , and .
Why responses add up as a convolution
Consider starting from rest. A short burst of forcing at time , of size , is like a small impulse. The response to a unit impulse at time is , so a burst at time produces afterward. The equation is linear, so the responses to all the bursts add, and adding them is the integral
In transform terms, , which is the convolution theorem.
Worked examples
Common mistakes
Practice problems
-
Compute .
Answer
Full solution
. Check: , the transform of .
-
Compute .
Answer
Full solution
.
-
Use convolution to find .
Answer
Full solution
.
-
Write the solution of , , as a convolution.
Answer
Full solution
and inverts to .
-
Using Exercise 4, solve from rest.
Answer
Full solution
.
-
Solve .
Answer
Full solution
The integral is . , so .
-
What is the impulse response of ?
Answer
Full solution
. Any forcing gives .
-
A student inverts as . Correct it.
Answer
The inverse is .
Full solution
A product of transforms corresponds to a convolution, and , which matches .
Frequently asked questions
What is a convolution?
(f * g)(t) = ∫₀ᵗ f(τ) g(t − τ) dτ. It combines two functions by weighting each past value of f by g at the time elapsed since.
What is the convolution theorem?
L{f * g} = F(s) G(s): the transform of a convolution is the product of the transforms.
Is convolution commutative?
Yes. f * g = g * f, as the substitution u = t − τ shows.
What is the impulse response?
The solution h(t) with zero initial conditions and forcing δ(t). For any forcing g, the solution is h * g.
Does the transform of fg equal FG?
No. The product FG comes from the convolution f * g, not from the ordinary product of the functions.