Mean and variance, and link to Poisson processes

Mean and variance, and link to Poisson processes

Spec SF3–SF4 · AQA A-Level Further Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • E(X)=1λ\mathrm{E}(X)=\frac1\lambda, Var(X)=1λ2\mathrm{Var}(X)=\frac1{\lambda^2} and σ=1λ\sigma=\frac1\lambda.
  • The proofs use integration by parts; state that the boundary term is 00.
  • In a Poisson process, counts in time tt are Po(λt)\mathrm{Po}(\lambda t) and gaps are exponential with parameter λ\lambda.
  • P(gap>t)=P(no events in t)=e−λt\mathrm{P}(\text{gap}>t)=\mathrm{P}(\text{no events in }t)=e^{-\lambda t}.

Mean, variance and standard deviation

The mean and standard deviation are both 1λ\frac1\lambda; the variance is 1λ2\frac1{\lambda^2}.

For XX exponential with parameter λ\lambda:

E(X)=1λ,Var(X)=1λ2,σ=1λ.\mathrm{E}(X)=\frac{1}{\lambda},\qquad \mathrm{Var}(X)=\frac{1}{\lambda^2},\qquad \sigma=\frac{1}{\lambda}.

The mean and standard deviation are equal. A higher rate λ\lambda means shorter gaps on average.

0.511.522.533.544.550.511.522.53xyf(x) = λe^(−λx)
The exponential probability density function f(x) = λe^(−λx): a larger rate λ gives shorter gaps on average, a mean of 1/λ
  • Mean1λ\dfrac1\lambda
  • Variance1λ2\dfrac1{\lambda^2}
  • Standard deviation1λ\dfrac1\lambda

Worked example

λ=0.05\lambda=0.05 per day. Find the mean, variance and standard deviation.

An exponential distribution has mean 4. What is its variance?

Proofs of the mean and variance

Integrate by parts; the boundary terms vanish because the exponential decays faster than xx grows.

Mean. E(X)=∫0∞xλe−λx dx\mathrm{E}(X)=\int_0^\infty x\lambda e^{-\lambda x}\,dx. By integration by parts with u=xu=x and v=−e−λxv=-e^{-\lambda x}:

[−xe−λx]0∞+∫0∞e−λx dx=0+1λ.\left[-xe^{-\lambda x}\right]_0^\infty+\int_0^\infty e^{-\lambda x}\,dx=0+\frac{1}{\lambda}.

Variance. By parts again, E(X2)=[−x2e−λx]0∞+∫0∞2xe−λx dx=2λE(X)=2λ2\mathrm{E}(X^2)=\left[-x^2e^{-\lambda x}\right]_0^\infty+\int_0^\infty2xe^{-\lambda x}\,dx=\frac{2}{\lambda}\mathrm{E}(X)=\frac{2}{\lambda^2}.

State the boundary term is zero because e−λxe^{-\lambda x} decays faster than xx grows.

  1. 1

    Integrate by parts

    u = x, v = −e^(−λx)

  2. 2

    Boundary term

    zero at both ends

  3. 3

    Remaining integral

    gives 1/λ

  4. 4

    E(X²)

    by parts again, reusing E(X)

  5. 5

    Subtract

    Var = 2/λ² − 1/λ² = 1/λ²

Proof outline

Which expression gives E(X2)\mathrm{E}(X^2) for an exponential distribution?

Linking the two distributions

A gap longer than tt is the same event as no events in time tt, with probability e−λte^{-\lambda t}.

A gap longer than tt is the same event as no events in a time tt. Both have probability e−λte^{-\lambda t}:

P(X>t)=P(Poisson with mean λt equals 0)=e−λt.\mathrm{P}(X>t)=\mathrm{P}(\text{Poisson with mean }\lambda t\text{ equals }0)=e^{-\lambda t}.

Use the Poisson count for "how many events in a period", and the exponential for "how long until the next event".

If the rate varies, for instance faults are more common near joints, neither model applies.

0.511.522.533.544.550.20.40.60.81xyP(X > t) = e^(−λt)
P(X > t) = e^(−λt): the chance of a gap longer than t, which is the chance of no events in time t
  • Gap longer than ttno events in time tt
  • Probabilitye−λte^{-\lambda t}

Worked example

Faults occur at λ=0.8\lambda=0.8 per km. Find the probability that the gap between faults is longer than 2 km.

Events form a Poisson process with rate λ\lambda. What is the probability of no events in time tt?

Try an exam question

Faults occur randomly along a cable at a constant average rate of 0.80.8 per km. Find the probability that the distance between two successive faults is more than 22 km.

[3 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Find the mean lifetime of a battery.

1 mark, marked against the mark scheme

Revision notes