Let’s assume that there is some number x in between 1 and k (1 … x … k). x is the minimum of the Xi’s. We need to find out the probavility that Y equals x.
Since we don’t know what exactly variable has minimum in Y we have to figure out the number of all posible ways that X1…Xn can be assign to 1…k. The possible number of assignments is \(k^n\) (k numbers to choose from and we chose n of them).
Next we have to count the number of ways that X1…Xn can be assign to x…k. The possible number of assignments is \((k-x)^n\) (k-x numbers to choose from and we chose n of them).
Now we have to count the number of ways that X1…Xn can be assign to x-1…k (the range includes x). The possible number of assignments is \((k-(x-1))^n\) (k-x numbers to choose from and we chose n of them).
P(Y=x) = \(\frac{(k-(x-1))^n-(k-x)^n}{k^n}=\frac{(k-x+1)^n-(k-x)^n}{k^n}\)
It’s a case of geometric distribution.
Probability density function: \(P(X = x) = p(1 - p)^{x-1} = p(q)^{x-1}\),
where p is probability of success and q is probability of failure
Expected value: \(E(X) = 1/p\)
Variance: \(Var(X) = q/p^2\)
Standard deviation: \(\sigma = \sqrt{Var(X)}\)
#probability that the machine fails each year
p <- 1/10
# probability that machine continues to work each year
q <- 1-p
#expected value
exp_value <- 1/p
exp_value
## [1] 10
#mashine will fail after 8 years on 9th year
x <- 9
# probability that the machine will fail after 8 years
prob <- p * q^(x-1)
prob
## [1] 0.04304672
# What is the Expected Value?
expected_value <- 1/p
expected_value
## [1] 10
#Variance
variance <- q/p^2
#Standard Deviation
sd <- sqrt(variance)
sd
## [1] 9.486833
Exponential Cumulative Distribution is defined by the following formula:
\(X \leq k\) : \(P (X \leq k) = 1 - e^{-k/\mu}\)
\(X \geq k\) : \(P (X \geq k) = e^{-k/\mu}\), where \(\mu = 1/\lambda\)
Expected Value: \(E(X) = 1/\lambda\)
Standard Deviation: \(\sigma = \sqrt{1/\lambda^2}\)
#parameter
par <- 1/10
#expectation
mu <- 1/par
year <- 8
# probability that the machine will fail after 8 years
prob <- exp(-year/mu)
prob
## [1] 0.449329
#expected value
exp_value <- 1/par
exp_value
## [1] 10
#standard deviation
sd <- sqrt(1/(par^2))
sd
## [1] 10
Binomial Distribution Model: \(P(X="success") = {n \choose k} p^x(1-p)^{n-x}\) Expected Value: \(E(X) = np\) Standard Deviation: \(\sigma = \sqrt{np(1-p)}\)
# probability of machine failure per year
p <- 0.1
# number of years
n <- 8
# probability of 0 success in 8 years
prob <- choose(n, 0) * p^0 * (1-p)^(8)
prob
## [1] 0.4304672
#expected value
exp_value <- n * p
exp_value
## [1] 0.8
# standard deviation
sd <- sqrt(n*p*(1 - p))
sd
## [1] 0.8485281
Poisson Distribution Model: \(P(X = x) = \frac{e^{-\mu}\mu^x}{x!}\) where μ is the mean number of “successes” and x is the number of “successes” in question.
Expected Value: \(E(X) = \mu\)
Standard Deviation: \(\sigma = \sqrt{\mu}\)
#number of years = the number of “successes” in question
x <- 8
mu <- 10
# number of years
n <- 8
# probability that the machine will fail after 8 years
prob <- (exp(-1 * mu) * mu^x) / factorial(x)
prob
## [1] 0.112599
#expected value
exp_value <- mu
exp_value
## [1] 10
# standard deviation
sd <- sqrt(mu)
sd
## [1] 3.162278