From exercise 10, the minimum value of an exponential density of \(n\) independent random variables is \(\mu\)/n.
Here \(\mu\) = 1000 hours. n=100
mu = 1000
n = 100
(excected_hours = mu/n)
## [1] 10
The PDF is convolution of two exponential is: \[ f_{X1}(x)= f_{X2}(x) = \begin{equation*} \left\{ \begin{array}{ll} \lambda e^{-\lambda x}, & \quad x > 0, \\ 0, & \quad otherwise. \end{array} \right. \end{equation*} \]
The convolution of two exponential densities if \(Z = X_1 - X_2\) is: \[ f_Z(z)=\int_{-\infty }^{\infty } f_{X_2}(x_1-z)f_{X_1}(x_1)dx_1 =\int_{0 }^{\infty } \lambda e^{-\lambda (x_1-z)}\lambda e^{-\lambda x_1} = \int_{0 }^{\infty } \lambda^2 e^{\lambda (z-2x_1)}\ \] Going to https://www.integral-calculator.com/ to calculate the integral gets:
\[ \frac{1}{2}\lambda e^{-\lambda z} \]
If for any positive number = \(k\sigma = k\) standard deviations for some integer \(k\), then
\[ P(|X-\mu|\ge k\sigma) \le\frac{1}{k^2} \]
\[ k\sigma = 2 \] \[ k= \frac{2}{\sqrt{100/3}} \] \[ \frac{1}{k^2}= 1/(\frac{2}{\sqrt{100/3}})^2 \]
(1/(2^2/(100/3)))
## [1] 8.333333
Here the upper bound is greater than 1 which is not possible therefore the answer is 1.
\[ k\sigma = 5 \]
\[ k= \frac{5}{\sqrt{100/3}} \] \[ \frac{1}{k^2}= 1/(\frac{5}{\sqrt{100/3}})^2 \]
(1/(5^2/(100/3)))
## [1] 1.333333
Here the upper bound is greater than 1 which is not possible therefore the answer is 1.
\[ k\sigma = 9 \]
\[ k= \frac{9}{\sqrt{100/3}} \] \[ \frac{1}{k^2}= 1/(\frac{9}{\sqrt{100/3}})^2 \]
(1/(9^2/(100/3)))
## [1] 0.4115226
\[ k\sigma = 20 \]
\[ k= \frac{20}{\sqrt{100/3}} \] \[ \frac{1}{k^2}= 1/(\frac{20}{\sqrt{100/3}})^2 \]
(1/(20^2/(100/3)))
## [1] 0.08333333