We know there are N total observations, that there are k unique values for each of these observations and that Y is the minimum.
(k-Y+1)^n represents the total number of permutations of X values, where the X values are between and include Y and k. This alone does not guarantee that any value of X is equal to Y.
(k-Y)^n represents the total number of permutations of X values where all X values are greater than Y.
By subtracting the first term above, from the second, we can calculate the number of permutations where at least one value of X is equal to Y. We then can divide this by our total number of permutations X to calculate the PDF.
P(min=Y) = ((k-Y+1)^n - (k-Y)^n )/k^n
If we expect one failure every 10 years, using a geometric distribution, we know that the probability of failure in any give year is (1/10). We also know that the probability of the machine not failing in a given year is (9/10).
As the question does not specify in what year the machine fails, we effectively want to calculate the likelihood of 8 consecutive years where the machine does not fail.
(9/10)^8
## [1] 0.4304672
The expected value is equal to 1/p:
1/(1/10)
## [1] 10
The Standard Deviation is equal to [(1-p)/p^2]^(1/2)
((1-.1)/(.1^2))^.5
## [1] 9.486833
We can calculate the probability the machine will fail after 8 years by taking e to the ratio of the years before failure * -1, over the total expected lifespan of the machine.
e^(-YearsBeforeFailure/Expected Lifespan)
exp(-8/10)
## [1] 0.449329
1/(1/10)
## [1] 10
(1/((1/10)^2))^.5
## [1] 10
This will look a similar to our geometric distribution. The probability of a “success” is .1, and a failure ,9. 8 consecutive failures can be calculated by:
.9^8
## [1] 0.4304672
e(x) = np
8*.1
## [1] 0.8
SD = (p(1-p)n)^.5
(.1*(1-.1)*8)^.5
## [1] 0.8485281
For this one we will use Rs Poisson functions
ppois(0,lambda = (8/10))
## [1] 0.449329
e(x)= lambda
8/10
SD = lambda^.5
(8/10)^.5
## [1] 0.8944272