11 21 computing a pmf px x collect all possible

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ust be TAIL) = P (TTTT ....H ) (tosses are independent) = (1 − p )k −1 p , M. Chen (IE@CUHK) ENGG2430C lecture 5 k = 1, 2, .... 11 / 21 Computing a PMF pX (x ) Collect all possible outcomes ω ∈ Ω, such that X (ω ) = x : {ω : ω ∈ Ω, X (ω ) = x } pX (x ) = P ({ω : ω ∈ Ω, X (ω ) = x }). Repeat for all x ∈ R . Example: two independent throws of a fair die. Define X and Y as the outcomes of the first and second throw, respectively 9 Define Z = min(X ;Y ). pZ (2) = 36 M. Chen (IE@CUHK) ENGG2430C lecture 5 12 / 21 Binomial Random Variable Experiment: flip a coin Independently for n times. Define Bernoulli/indicator r.v.s Xi , (i = 1, 2, . . . , n) be the outcome of the i -th flip: 1, the ith toss gives HEAD; Xi = 0, the ith toss gives TAIL. The PMF is pXi (1) = p and pXi (0) = 1 − p . Let Y = ∑n=1 Xi . Then Y is a Binomial random variable. Its PMF is i given by n pY (k ) = P (Y = k ) = P ∑ Xi = k i =1 = (n ) p k (1 − p )n−k . k M. Chen (IE@CUHK) ENGG2430C lecture 5 13 / 21 Service Facility Design n: the number of customers p : probability that a customer requires service s : number of service lines X : a r.v. denoting the number of service requests Goal: Design a system that “no-service" probability is no more than ε : P (X > s ) = PX (s + 1) + · · · + pX (n) n n = ∑ k p k (1 − p )n−k . k =s +1 The larger the s , the smaller the “no-service" probability is. However, it costs more to run the service. Given ε , the optimal s ∗ (ε ) = min(s ∈ N : P (X > s ) ≤ ε ). M. Chen (IE@CUHK) ENGG2430C lecture 5 14 / 21 Expectation The so-called first-order statistics about random outcomes: E [X ] = ∑ xpX (x ). (A weighted sum of outcomes x ’s.) x Interpretations: The center of gravity of the probability mass The average of repeated experiments Example: a sample space Ω = {0, 1, . . . , n} with a uniform probability law (From MIT open course 6.041 / 6.431 lecture notes.) E [X ] = 0 × M. Chen (IE@CUHK) 1 1 n n +1× +···+ = n+1 n+1 n+1 2 ENGG2430C lecture 5 15 / 21 Examples Calculate the expectation of a Bernoulli random...
View Full Document

Ask a homework question - tutors are online