Why does the sum of two uniforms give a triangle?

Convolving two rectangular (uniform) densities produces a triangular density. Geometrically, as you slide one rectangle across the other, the overlap area first increases linearly (rising edge) then decreases linearly (falling edge), tracing a triangle shape.

What is the connection between convolution and the Central Limit Theorem?

The CLT says that sums of many independent variables approach a Gaussian. Since sums correspond to convolutions of densities, the CLT means repeated self-convolution of any density converges to a Gaussian shape. This is the convolution-theoretic statement of the CLT.

How do moment generating functions simplify convolution?

MGFs convert convolution to multiplication: M_{X+Y}(t) = M_X(t)·M_Y(t). This parallels how Laplace transforms convert time-domain convolution to s-domain multiplication. Find individual MGFs, multiply, then identify the result — often much easier than evaluating the convolution integral directly.

Convolution in Probability

Quick Answer

In probability theory, the convolution of two probability density functions f_X and f_Y gives the density of their sum Z = X + Y (assuming independence): f_Z(z) = ∫f_X(x)·f_Y(z−x)dx. For example, the sum of two independent uniform(0,1) variables has a triangular distribution on (0,2), obtained by convolving the rectangular density with itself. This extends to any number of independent variables: the distribution of X₁ + X₂ + ... + Xₙ is the n-fold convolution of their individual densities.

Calculate this instantly on LAPLACE Calculator →

Why Sums of Random Variables Involve Convolution

If X and Y are independent random variables and Z = X + Y, then P(Z ≤ z) = P(X + Y ≤ z). To find this probability, integrate over all pairs (x, y) where x + y ≤ z, weighted by the joint density f_X(x)·f_Y(y) (which factors due to independence). Changing variables to x and z−x and integrating over x gives f_Z(z) = ∫f_X(x)·f_Y(z−x)dx — which is exactly the convolution integral. The flip-and-slide interpretation makes physical sense: for each possible value of X = x, the remaining amount z − x must come from Y, and you sum over all such partitions.

Key Formulas

Classic Examples: Uniform, Exponential, and Normal Sums

Convolving two Uniform(0,1) densities (rectangles of height 1, width 1) produces a triangle on (0,2) with peak at z=1. Convolving an Exponential(λ) with itself gives a Gamma(2,λ) distribution — the Erlang distribution describing the time for two Poisson events. Convolving n exponentials gives Gamma(n,λ). The most elegant result: convolving two Gaussians N(μ₁,σ₁²) and N(μ₂,σ₂²) yields another Gaussian N(μ₁+μ₂, σ₁²+σ₂²). Means add and variances add — the Gaussian family is closed under convolution, a property shared by no other common distribution family.

Compute convolution in probability Instantly

Get step-by-step solutions with AI-powered explanations. Free for basic computations.

Open Calculator

The Central Limit Theorem as Repeated Convolution

The central limit theorem states that the sum of many independent random variables approaches a Gaussian distribution regardless of the individual distributions. In convolution language: repeatedly convolving any density function with itself drives the result toward the Gaussian shape. After just 3-4 self-convolutions, most distributions already look nearly Gaussian. This happens because convolution smooths and spreads distributions, and the Gaussian is the unique fixed point of the convolution-and-rescaling operation. Mathematically, the Fourier transform of a convolution is a product, and repeated multiplication of characteristic functions converges to the Gaussian characteristic function.

Moment Generating Functions: Convolution Made Easy

The moment generating function (MGF) M_X(t) = E[e^{tX}] transforms convolution into multiplication, just like the Laplace transform. If Z = X + Y (independent), then M_Z(t) = M_X(t)·M_Y(t). This is because E[e^{t(X+Y)}] = E[e^{tX}]·E[e^{tY}] by independence. The MGF is essentially the bilateral Laplace transform of the density function. Similarly, the characteristic function φ_X(t) = E[e^{itX}] (the Fourier transform of the density) converts convolution to multiplication: φ_Z = φ_X · φ_Y. These transforms provide the fastest route to finding the distribution of sums.

Applications: Insurance, Queuing, and Risk Analysis

Insurance companies model total claims as the sum of individual claim amounts — convolution of claim size distributions. If individual claims follow an exponential distribution, total monthly claims follow a Gamma distribution (n-fold exponential convolution). Queuing theory models total service time as the convolution of individual service time distributions. Financial risk analysis uses convolution to model portfolio returns as the sum of individual asset returns. In each case, the convolution operation propagates uncertainty from individual components to the aggregate, enabling probability calculations about totals, maxima, and threshold crossings.

Frequently Asked Questions

Convolution of two probability density functions gives the distribution of the sum of two independent random variables. If Z = X + Y, then f_Z = f_X * f_Y (the convolution of the densities of X and Y).

Master Your Engineering Math

Join thousands of students and engineers using LAPLACE Calculator for instant, step-by-step solutions.

Start Calculating Free →