Inverse poisson distribution calculation

2.9k Views Asked by At

Previous studies show that there is an average of 0.2 breakdowns on each on kilometre of the Wellington motorway on any particular day.

If a towing service wants to be available to tow vehicle that has broken down, what length of the motor way would they have to patrol to have a probability of 0.75 of being close to at least one breakdown.

Confusion

Since we are given the average (0.2) intuitively it almost feels like I can do this

$$ \frac{0.75}{0.2} . 1 = 3.75km$$

as if we double the amount of kilometres then we double the probability. We can't and I don't understand why. Because it seems like a rectangular distribution where the probability is the same throughout each kilometre.

1

There are 1 best solutions below

0
On BEST ANSWER

Another way to frame the question is to say that, on average, a towing service would need to patrol $5$ kilometers to see one breakdown. The random number of breakdowns $X$ observed in $k$ kilometers is modeled as a Poisson distribution with rate $\lambda = 0.2k$; that is to say, $$\Pr[X = x] = e^{-\lambda} \frac{\lambda^x}{x!}, \quad x = 0, 1, 2, \ldots.$$ So the probability of observing at least one breakdown in $k$ kilometers of patrolled road is $$\Pr[X \ge 1] = 1 - \Pr[X = 0] = 1 - e^{-\lambda} = 1 - e^{-0.2k}.$$ All that's left is to note that we want to find the smallest $k$ for which this probability is at least $0.75$; i.e., to solve $$1 - e^{-0.2k} = 0.75.$$

The reason why your solution doesn't work is because if $0.75$ were replaced with $1$, then it would mean that you would be guaranteed to find a breakdown within a $5$ kilometer patrol region. But this is not what the question stipulates. Rather, it simply says that on average, there is one breakdown in a given $5$ kilometer stretch of road.