probability generating functions

Probability Generating Functions (PGFs) are powerful analytical tools for discrete random variables, providing a compact representation of probability distributions.

Definition: For a discrete random variable X taking non-negative integer values, the PGF is defined as:

$$G_X(t) = E(t^X) = \sum_{x=0}^{\infty} P(X=x)t^x$$

where t is a dummy variable and |t| ≤ 1 for convergence.

Key Properties:

1. Probability Recovery: $P(X=r) = \frac{1}{r!}\left[\frac{d^r G_X}{dt^r}\right]_{t=0}$
2. Normalisation: $G_X(1) = 1$ (sum of all probabilities)
3. Expectation: $E(X) = G_X'(1)$ (first derivative at t = 1)
4. Variance: $\text{Var}(X) = G_X''(1) + G_X'(1) - [G_X'(1)]^2$
5. Sum of Independent Variables: If X and Y are independent, $G_{X+Y}(t) = G_X(t) \cdot G_Y(t)$

Standard PGFs you must memorise:

• Binomial B(n, p): $G(t) = (q + pt)^n$ where q = 1 - p
• Poisson Po(λ): $G(t) = e^{\lambda(t-1)}$
• Geometric Geo(p): $G(t) = \frac{pt}{1-qt}$

Mnemonic: "PGFs Generate Probabilities" — remember the Generating function Generates all distribution properties.

Cambridge Note: PGFs are particularly valuable for finding distributions of sums and understanding composite probability scenarios — skills directly tested in Paper 3.

Think of a PGF as a probability-encoding machine: each coefficient of t^r holds the probability that X = r. The function "generates" all probabilities simultaneously.

Real-World Application 1: Quality Control

A manufacturing plant produces batches where defective items follow a Poisson distribution Po(2.5). Using PGFs, quality controllers can:

• Calculate the distribution of total defects across multiple batches
• Determine probability of zero defects using $G(0) = P(X=0)$
• Find expected defects using $G'(1) = \lambda = 2.5$

Real-World Application 2: Network Reliability

In computer networks, packet loss follows geometric distributions. If server A has Geo(0.7) loss and independent server B has Geo(0.6) loss, the PGF of total losses is:

$$G_{\text{total}}(t) = \frac{0.7t}{1-0.3t} \times \frac{0.6t}{1-0.4t}$$

This multiplication property makes PGFs invaluable for system reliability analysis.

Analogy: Imagine a PGF as a musical chord: just as a chord encodes multiple notes simultaneously, a PGF encodes all probabilities at once. Differentiation "extracts" specific notes (probabilities), while setting t = 1 plays the "full chord" (total probability).

Why PGFs Matter:

• They transform addition of random variables into multiplication of functions
• Complex probability calculations become algebraic manipulations
• They reveal structural properties invisible in probability tables

Practical Insight: PGFs excel when dealing with sums of independent variables — Cambridge examiners frequently test this property.