TheoremComplete

The Convolution Theorem

The convolution theorem establishes a fundamental connection between multiplication in the transform domain and convolution in the time domain.


Statement

Theorem4.11Convolution theorem

If L{f}=F\mathcal{L}\{f\} = F and L{g}=G\mathcal{L}\{g\} = G, then

L{fg}(s)=F(s)G(s)\mathcal{L}\{f * g\}(s) = F(s)G(s)

where the convolution is (fg)(t)=0tf(τ)g(tτ)dτ(f * g)(t) = \int_0^t f(\tau)g(t - \tau)\,d\tau.


Proof

Proof

F(s)G(s)=0esτf(τ)dτ0esσg(σ)dσ=00es(τ+σ)f(τ)g(σ)dτdσ.F(s)G(s) = \int_0^\infty e^{-s\tau}f(\tau)\,d\tau \cdot \int_0^\infty e^{-s\sigma}g(\sigma)\,d\sigma = \int_0^\infty\int_0^\infty e^{-s(\tau+\sigma)}f(\tau)g(\sigma)\,d\tau\,d\sigma.

Substitute t=τ+σt = \tau + \sigma (so σ=tτ\sigma = t - \tau, dσ=dtd\sigma = dt). The region of integration τ0,σ0\tau \geq 0, \sigma \geq 0 becomes 0τt0 \leq \tau \leq t, t0t \geq 0:

F(s)G(s)=0est(0tf(τ)g(tτ)dτ)dt=0est(fg)(t)dt=L{fg}(s).F(s)G(s) = \int_0^\infty e^{-st}\left(\int_0^t f(\tau)g(t-\tau)\,d\tau\right)dt = \int_0^\infty e^{-st}(f*g)(t)\,dt = \mathcal{L}\{f*g\}(s).

The interchange of order of integration is justified by absolute convergence (exponential order condition). \blacksquare


Applications

ExampleSolving integral equations

Solve the integral equation y(t)=t+0ty(τ)sin(tτ)dτy(t) = t + \int_0^t y(\tau)\sin(t-\tau)\,d\tau.

This is y=t+ysinty = t + y * \sin t. Taking transforms: Y=1s2+Y1s2+1Y = \frac{1}{s^2} + Y \cdot \frac{1}{s^2+1}.

Y(11s2+1)=1s2    Ys2s2+1=1s2    Y=s2+1s4=1s2+1s4.Y\left(1 - \frac{1}{s^2+1}\right) = \frac{1}{s^2} \implies Y\cdot\frac{s^2}{s^2+1} = \frac{1}{s^2} \implies Y = \frac{s^2+1}{s^4} = \frac{1}{s^2} + \frac{1}{s^4}.

Inverting: y(t)=t+t36y(t) = t + \frac{t^3}{6}.

ExampleVariation of parameters via convolution

The solution to y+ω2y=g(t)y'' + \omega^2 y = g(t), y(0)=y(0)=0y(0) = y'(0) = 0 is:

Y(s)=G(s)s2+ω2=G(s)1s2+ω2.Y(s) = \frac{G(s)}{s^2+\omega^2} = G(s) \cdot \frac{1}{s^2+\omega^2}.

By the convolution theorem: y(t)=1ω0tg(τ)sin[ω(tτ)]dτy(t) = \frac{1}{\omega}\int_0^t g(\tau)\sin[\omega(t-\tau)]\,d\tau.

This is precisely the variation of parameters formula, derived algebraically.


Properties of Convolution

RemarkAlgebraic properties

Convolution satisfies:

  1. Commutativity: fg=gff * g = g * f.
  2. Associativity: (fg)h=f(gh)(f * g) * h = f * (g * h).
  3. Distributivity: f(g+h)=fg+fhf * (g + h) = f * g + f * h.
  4. Zero element: f0=0f * 0 = 0.
  5. No identity: There is no ordinary function ee with fe=ff * e = f for all ff. However, the Dirac delta δ\delta serves as a generalized identity: fδ=ff * \delta = f.
ExampleDirect computation

Compute (ete2t)(t)=0teτe2(tτ)dτ=e2t0teτdτ=e2t(1et)=e2tet(e^t * e^{2t})(t) = \int_0^t e^\tau e^{2(t-\tau)}\,d\tau = e^{2t}\int_0^t e^{-\tau}\,d\tau = e^{2t}(1-e^{-t}) = e^{2t} - e^t.

Verification: L{e2tet}=1s21s1=1(s1)(s2)=1s11s2=F(s)G(s)\mathcal{L}\{e^{2t}-e^t\} = \frac{1}{s-2}-\frac{1}{s-1} = \frac{1}{(s-1)(s-2)} = \frac{1}{s-1}\cdot\frac{1}{s-2} = F(s)G(s).