The minimum of the Xi’s (X1, X2, . . . , Xn) is denoted by Y, and each is uniformly distributed from 1 to k => each can be assumed to have k possibilities, therefore to find the distribution function P(X=m) of Y, I have to: - deduce the number of possible ways to assign the Xi’s to the values between m and k with at least one Xi being assigned to m - And deduce the total number of combinations to assign the Xi’s to values between 1 and k, that is, kn.
$k^n = $ Total combinations. $(k−1)^n = $ combinations where none of the Xi’s is equal to 1.
=> $ P(X=1) = $
Similarly:
\(P(X=2) = \frac{(k−2+1)^n−(k−2)^n}{k^n}\)
\(P(X=3) = \frac{(k−3+1)^n−(k−3)^n}{k^n}\)
Generalizing this steps for the distribution function (X=m), then:
\(P(X=m) = \frac{(k−m+1)^n−(k−m)^n}{k^n}\)
sim <- function(k,n,trials=100000) {
Y<-rep(0,trials)
for (i in 1:trials) {
x<-sample.int(k,size=n,replace=TRUE)
Y[i]<-min(x)
}
return(Y)
}
par(mfrow=c(1,2))
k<-100
n<-20
hist(sim(k,n),breaks=60,main=paste("Simulated: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
pY<-((k-1:k+1)^n-(k-1:k)^n)/k^n
barplot(pY,main=paste("Theoretical: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
par(mfrow=c(1,2))
k<-100
n<-5
hist(sim(k,n),breaks=60,main=paste("Simulated: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
pY<-((k-1:k+1)^n-(k-1:k)^n)/k^n
barplot(pY,main=paste("Theoretical: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
par(mfrow=c(1,2))
k<-20
n<-5
hist(sim(k,n),breaks=60,main=paste("Simulated: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
pY<-((k-1:k+1)^n-(k-1:k)^n)/k^n
barplot(pY,main=paste("Theoretical: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
par(mfrow=c(1,2))
k<-20
n<-100
hist(sim(k,n),breaks=60,main=paste("Simulated: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
pY<-((k-1:k+1)^n-(k-1:k)^n)/k^n
barplot(pY,main=paste("Theoretical: k=",k," and n=",n,sep=""),xlab="Y",xlim=c(1,k))
This is a geometric distribution.
Since one failure occurs every ten years => \(p=0.1 => q =1−p=0.9\)
\(F_x(k) = P(X≤k)=1−q^{k+1}\): k = number of failures before the first success. In other words, \(P(X>k)=1−P(X≤k)=1−(1−q^{k+1})=qk+1\). When k=8 => \(P(X>8)=0.9^9 = 0.3874\).
Using R geometric distribution function
pgeom(8, 0.1, lower.tail=FALSE)
## [1] 0.3874205
Therefore, The expected number of years before the first failure is \(E(X)=\frac{q}{p}=\frac{0.9}{0.1}=9years\)
Standard deviation \(\sigma^2=\sqrt{\frac{q}{p^2}} = \sqrt{\frac{0.9}{0.12}} \approx{9.4868}\)
For the exponential distribution, CDF:
\(FX(k)=P(X≤k)=1−e^{−λk}\), where λ = the rate parameter. We are given that λ=0.1 => \(P(X>k)=1−P(X≤k)=1−(1−e^{−λk})=e^{−λk}\) When k=8 => \(P(X>8)=e^{−0.8} \approx{0.4493}\) This can also be verified with R as follows:
pexp(8, 0.1, lower.tail=FALSE)
## [1] 0.449329
Therefore expected value: \(E(X)=\frac{1}{λ}=\frac{1}{0.1}=10\).
Standard deviation: \(\sigma^2=\sqrt{\frac{1}{λ^2}} = \frac{1}{λ}=10\).
Therefore, \(P(X=k)=\binom{n}{k}p^kq^{n−k}\). As noted before, the Probability of a failure after 8 years = the probability of 0 successes after 8 trials. When k=0 and n=8 => \(P(X=0)=\binom{8}{0}0.1^0×0.9^{8−0}=1×1×0.9^8 = 0.4305\)
The expected value E(X) and will depend on number of years tracked.
For the first 8 years:
\(E(X)=np=8×0.1=0.8\)
Then
\(\sigma = \sqrt{npq} = \sqrt{8×0.1×0.9} = 0.8485\).
Since average number of failures in every 10 years is 1 so average number of failures in 8 years will be:
\(λ=\frac{8}{10}=0.8\)
Let X = failures in 8 years. The probability of 0 failures in 8 years will be =>
\(P(X=0)=(\frac{λ^k}{k!}).e^{−λ} = (\frac{0.8^0}{0!}).e^{−0.8}=e^{−0.8}\)
=> \(P(X=0)=0.4493\)
ppois(0,0.1,lower.tail=TRUE)^8
## [1] 0.449329
E(X): \(λ=0.8\)
\(\sigma = \sqrt{0.8} = 0.8944\) failures