(ELEC2600) (2021) (S) Final 7em935 25400
(ELEC2600) (2021) (S) Final 7em935 25400
2. (14 pts) Let 𝑋 be the lifetime (measured in years) of a component in a system, which is
continuous and uniformly distributed in the interval 7, 10 . The system also has a backup
component that can take over when the original component fails so that the system can provide
continuous service. Let 𝑌 be the lifetime of the backup component, which is an exponential
random variable with mean 7 years. Assume that the lifetimes of the original and backup
components are independent. Let 𝑍 𝑋 𝑌 be the total lifetime of the system.
(a) Find the mean of 𝑍. (2 pts)
(b) Find the pdf of 𝑍. Note that you will need different expressions for different ranges of 𝑍.
For partial credit, give the ranges of 𝑍. For full credit, find the correct expressions in each
of these ranges. (10 pts)
(c) Find the probability that the backup component lasts for between 6.5 and 7.5 years, i.e.
𝑃 6.5 𝑌 7.5 . (2 pts)
3. (16 pts) Suppose that each second, the number of requests to a web server, 𝑋 , 𝑖 0, 1, 2, … ,
is a discrete random variable that is uniformly distributed over the set A, where 𝐴 5, 7 .
Assume that the number of requests to the web serval at each second is independent from each
other. Let 𝑆 be the total number of requests to a web server in 𝑛 seconds. Assume that 𝑆 0.
It turns out that 𝑆 is a sum process.
(a) Find 𝐸 𝑋 𝑋 . (2 pts)
(b) Find 𝐸 𝑆 . (2 pts)
(c) Find Var 𝑆 . (3 pts)
(d) Find 𝐶 10, 20 𝐶𝑜𝑣 𝑆 , 𝑆 . (2 pts)
(e) Is 𝑆 a wide-sense stationary process? Justify your answer (2 pts)
(f) Use the central limit theorem to approximate 𝑃 𝑆 550 . Express your answer in terms
of the Q function, Q 𝑥 𝑃 𝑋 𝑥 , where 𝑋 is a Gaussian random variable with zero
mean and variance 1. (5 pts)
(ELEC2600)[2021](s)final~=7em935^_25400.pdf downloaded by bbisht from http://petergao.net/ustpastpaper/down.php?course=ELEC2600&id=12 at 2024-02-01 10:20:27. Academic use within HKUST only.
4. (16 pts) A Gaussian random process is one where all finite order distributions are jointly
Gaussian distributed. Suppose that 𝑋 𝑡 is a continuous-time Gaussian random process with
mean and covariance functions given by
𝑚 𝑡 𝐸𝑋 𝑡 cos and 𝐶 𝑡 , 𝑡 0.2𝑒 | |
.
(a) Is this process stationary? Why or why not? (2 pts)
(b) Write the variance function for 𝑋 𝑡 , Var 𝑋 𝑡 , for all ∞ 𝑡 ∞. (1 pt)
(c) Give the first order pdf for 𝑋 1 , 𝑓 𝑥 , for ∞ 𝑥 ∞. (3 pts)
(d) Define a random vector 𝑋⃗ 𝑋 1 𝑋 4 . Give the second order pdf for 𝑋⃗ , 𝑓 ⃗ 𝑥⃗
𝑓 𝑥 , 𝑥 where 𝑥⃗ 𝑥 ,𝑥 for ∞ 𝑥 , 𝑥 ∞. (4 pts)
(e) Find the correlation coefficient between 𝑋 1 and 𝑋 4 . (2 pts)
(f) Define a random variable 𝑍 7𝑋 1 6𝑋 4 . Give the pdf of 𝑍, 𝑓 𝑧 for ∞ 𝑧 ∞.
(4 pts)
5. (10 pts) Suppose that the total number of requests to a web server received between time 0 and
time t, N(t), is given by a Poisson random process with rate λ = 16 requests per minute. Assume
that time t is measured in minutes.
(a) Give the value of 𝐸 𝑁 11.5 . (1 pt)
(b) Give the value of Var 𝑁 11.5 . (1 pt)
(c) Give the value of 𝐶 11.5, 12.5 Cov 𝑁 11.5 , 𝑁 12.5 . (1 pt)
(d) Find the probability that exactly 17 requests are received in the first minute. (1 pt)
(e) Find 𝑃 𝑁 1 17 ∩ 𝑁 3 20 . (2 pts)
(f) Find the probability that the first request occurs before t = 2 seconds. (2 pts)
(g) Find the probability that the first request occurs between t = 3 and t = 6 seconds. (2 pts)