3 Expectation and variance of the poisson distribution
The expectation and variance of the Poisson distribution can be derived directly from the definitions which apply to any discrete probability distribution. However, the algebra involved is a little lengthy. Instead we derive them from the binomial distribution from which the Poisson distribution is derived.
Intuitive Explanation
One way of deriving the mean and variance of the Poisson distribution is to consider the behaviour of the binomial distribution under the following conditions:
- is large
- is small
- (a constant)
Recalling that the expectation and variance of the binomial distribution are given by the results
it is reasonable to assert that condition (2) implies, since , that is approximately 1 and so the expectation and variance are given by
In fact the algebraic derivation of the expectation and variance of the Poisson distribution shows that these results are in fact exact .
Note that the expectation and the variance are equal.
Key Point 9
The Poisson Distribution
If is the random variable {number of occurrences in a given interval}
for which the average rate of occurrences is and can assume the values and the probability of occurrences in that interval is given by
then the expectation and variance of the distribution are given by the formulae
For a Poisson distribution the Expectation and Variance are equal.
Exercises
-
Large sheets of metal have faults in random positions but on average have 1 fault per
.
What is the probability that a sheet 5 m 8 m will have at most one fault?
- If 250 litres of water are known to be polluted with bacteria what is the probability that a sample of 1 cc of the water contains no bacteria?
-
Suppose vehicles arrive at a signalised road intersection at an average rate of 360 per
hour and the cycle of the traffic lights is set at 40 seconds. In what percentage of cycles
will the number of vehicles arriving be
- exactly 5,
- less than 5? If, after the lights change to green, there is time to clear only 5 vehicles before the signal changes to red again, what is the probability that waiting vehicles are not cleared in one cycle?
-
Previous results indicate that 1 in 1000 transistors are defective on average.
- Find the probability that there are 4 defective transistors in a batch of 2000.
- What is the largest number, , of transistors that can be put in a box so that the probability of no defectives is at least 1/2?
- A manufacturer sells a certain article in batches of 5000. By agreement with a customer the following method of inspection is adopted: A sample of 100 items is drawn at random from each batch and inspected. If the sample contains 4 or fewer defective items, then the batch is accepted by the customer. If more than 4 defectives are found, every item in the batch is inspected. If inspection costs are 75 p per hundred articles, and the manufacturer normally produces 2 of defective articles, find the average inspection costs per batch.
-
A book containing 150 pages has 100 misprints. Find the probability that a particular page
contains
- no misprints,
- 5 misprints,
- at least 2 misprints,
- more than 1 misprint.
-
For a particular machine, the probability that it will break down within a week is 0.009. The
manufacturer has installed 800 machines over a wide area. Calculate the probability
that
- 5,
- 9,
- less than 5,
- more than 4 machines breakdown in a week.
-
At a given university, the probability that a member of staff is absent on any one day is 0.001.
If there are 800 members of staff, calculate the probabilities that the number absent on any
one day is
- 6,
- 4,
- 2,
- 0,
- less than 3,
- more than 1.
-
The number of failures occurring in a machine of a certain type in a year has a Poisson
distribution with mean 0.4. In a factory there are ten of these machines. What
is
- the expected total number of failures in the factory in a year?
- the probability that there are fewer than two failures in the factory in a year?
-
A factory uses tools of a particular type. From time to time failures in these tools occur and
they need to be replaced. The number of such failures in a day has a Poisson distribution with
mean 1.25. At the beginning of a particular day there are five replacement tools in stock. A
new delivery of replacements will arrive after four days. If all five spares are used before the
new delivery arrives then further replacements cannot be made until the delivery
arrives.
Find
- the probability that three replacements are required over the next four days.
- the expected number of replacements actually made over the next four days.
-
Poisson Process. In a sheet size
we expect
4 faults
-
In 1 cc
we expect
4 bacteria
-
In 40 seconds
we expect
4 vehicles
- (exactly 5) i.e. in 15.6 of cycles
- (less than 5)
Vehicles will not be cleared if more than 5 are waiting.
(greater than 5) (exactly 5) (less than 5)
-
-
Poisson approximation to binomial
-
-
Poisson approximation to binomial
-
(defective)
. Poisson approximation
to binomial
(4 or fewer defectives in sample of 100)
Inspection costs
Cost 75 0.947347 0.0526 -
- 0.51342
- 0.00056,
- 0.14430,
- 0.14430
-
- 0.12038,
- 0.10698,
- 0.15552,
- 0.84448
-
- 0.00016,
- 0.00767,
- 0.14379,
- 0.44933,
- 0.95258,
- 0.19121
-
Let
be the total number of failures.
-
Let the number required over 4 days be
Then
and
-
Let
be the number of replacements made.
and