The Poisson distribution

The Poisson distribution

Spec 2.1 · Edexcel A-Level Further Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • X∼Po(λ)X\sim\text{Po}(\lambda) counts events in a fixed interval: P(X=x)=e−λλxx!P(X=x)=\frac{e^{-\lambda}\lambda^x}{x!}, with no upper limit on xx.
  • The model needs events that occur singly, independently and at a constant mean rate.
  • P(X≥k)=1−P(X≤k−1)P(X\geq k)=1-P(X\leq k-1): get the boundary right.
  • The mean rate scales with the interval: tt times the rate gives Po(tλ)\text{Po}(t\lambda).
  • Independent Poisson variables add: X+Y∼Po(λ+μ)X+Y\sim\text{Po}(\lambda+\mu).

The Poisson model

The number of events in a fixed interval of time or space, when events occur singly, independently and at a constant mean rate.

The Poisson distribution models the number of events in a fixed interval of time or space. If X∼Po(λ)X\sim\text{Po}(\lambda), where λ>0\lambda>0 is the mean rate of events in the interval, then P(X=x)=e−λλxx!P(X=x)=\frac{e^{-\lambda}\lambda^x}{x!} for x=0,1,2,…x=0,1,2,\ldots

There is no upper limit on XX, and no fixed number of trials, unlike a binomial.

The model is suitable when events occur singly, independently of each other, at a constant mean rate. Examples: emails per 10 minutes, faults per metre of cloth, calls per hour.

00.050.10.150.20.00660.0023012345678910xP(X = x)
Po(3.5) probabilities: skewed, with the peak at x = 3.

Which situation is best modelled by a Poisson distribution?

Calculating probabilities

Use the calculator's Poisson and cumulative functions, and convert every inequality to P(X=x)P(X=x) or P(X≤x)P(X\leq x).

Use the Poisson function for P(X=x)P(X=x) and the cumulative probability function for P(X≤x)P(X\leq x). For X∼Po(3.5)X\sim\text{Po}(3.5):

  • P(X=2)=0.185P(X=2)=0.185
  • P(X≤2)=0.3208P(X\leq2)=0.3208
  • P(X>3)=1−P(X≤3)P(X>3)=1-P(X\leq3), which equals P(X≥4)P(X\geq4)
  • P(2<X≤6)=P(X≤6)−P(X≤2)=0.614P(2<X\leq6)=P(X\leq6)-P(X\leq2)=0.614

Writing P(X≥3)=1−P(X≤3)P(X\geq3)=1-P(X\leq3) is a common error: that is P(X>3)P(X>3).

00.050.10.150.20.00660.0023012345678910xP(X = x)
Probabilities for X ~ Po(3.5): P(X = 2) = 0.185. Summing the bars gives cumulative probabilities.
  • At least kkP(X≥k)=1−P(X≤k−1)P(X\geq k)=1-P(X\leq k-1)

Worked example

X∼Po(3.5)X\sim\text{Po}(3.5). Find P(X≥3)P(X\geq3).

X∼Po(3.5)X\sim\text{Po}(3.5). Which expression gives P(X≥5)P(X\geq5)?

Changing the interval

The mean rate scales with the length of the interval.

If events occur at a mean rate of λ\lambda per minute, the number in tt minutes is Po(tλ)\text{Po}(t\lambda).

For example, 6 calls per hour gives Po(2)\text{Po}(2) for 20 minutes and Po(1)\text{Po}(1) for 10 minutes. Always match the time unit in the question, and write the new distribution explicitly, such as X∼Po(2.5)X\sim\text{Po}(2.5), before using the calculator.

Separate periods are independent, so probabilities for two periods multiply. To count how many periods satisfy a condition, a binomial distribution may be used.

00.050.10.150.20.250.0120.00340.0009012345678xP(X = x)
Probabilities for X ~ Po(2), the number of calls in 20 minutes when the rate is 6 per hour: P(X = 1) = 0.2707.
  • Rate λ\lambda, interval ttX∼Po(tλ)X\sim\text{Po}(t\lambda)

Worked example

Calls arrive at a mean rate of 6 per hour. Find the probability of exactly 1 call in a 20-minute period, and of exactly 1 call in each of two separate 20-minute periods.

Cars pass a point at a mean rate of 12 per hour. What is the distribution of the number passing in 10 minutes?

The additive property

Independent Poisson variables add: their means add too.

If X∼Po(λ)X\sim\text{Po}(\lambda) and Y∼Po(μ)Y\sim\text{Po}(\mu) are independent, then X+Y∼Po(λ+μ)X+Y\sim\text{Po}(\lambda+\mu). No proof is required.

If online orders are Po(3)\text{Po}(3) and phone orders are Po(2)\text{Po}(2) per hour, independently, the total is Po(5)\text{Po}(5) per hour. Use the additive property to combine sources, then rescale for the interval needed.

The property needs independence. If the two sources are linked, the total is not Poisson in general, so state the assumption.

00.050.10.150.20.250.00990.00310.00090123456789xP(X = x)
Probabilities for the total orders in a 30-minute period, Po(2.5): P(X = 2) = 0.2565.
  • X∼Po(λ), Y∼Po(μ)X\sim\text{Po}(\lambda),\ Y\sim\text{Po}(\mu)X+Y∼Po(λ+μ)X+Y\sim\text{Po}(\lambda+\mu)

Worked example

Online orders are Po(3)\text{Po}(3) per hour and phone orders are Po(2)\text{Po}(2) per hour, independently. Find the probability of exactly 2 orders in total in a 30-minute period.

X∼Po(4)X\sim\text{Po}(4) and Y∼Po(1.5)Y\sim\text{Po}(1.5) are independent. What is the distribution of X+YX+Y?

Modelling and critical comment

Check each condition in context, name the one that fails and say how it affects the model.

To model a situation, choose λ\lambda as the observed mean rate for the interval.

To comment critically, check each condition in context: independence (does one event make another more or less likely?), constant rate (does the rate change within the period?) and singly (can events happen in groups?).

Name the condition that fails and explain in terms of the context. Also say what a probability means in context.

Independence

  • Fails: convoys, queues, epidemics

Constant rate

  • Fails: rush hour, evening peaks

Singly

  • Fails: a coach arriving with several passengers
  1. 1

    Name the condition

    independence, constant rate or events occurring singly

  2. 2

    Link to the context

    say why it might fail in this situation

  3. 3

    State the effect

    what happens to the model's probabilities

Structure of a critical comment.

Coaches arrive at a museum, each bringing several visitors at once. Which Poisson condition is most clearly broken for the number of visitors?

Try an exam question

Calls arrive at a switchboard at a mean rate of 6 per hour. Find the probability that in a 30-minute period there are (a) exactly 2 calls, (b) at least 4 calls.

[4 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Find P(X=2)P(X=2).

1 mark, marked against the mark scheme

Revision notes