Problem Set 1: Review of undergraduate probability.

Due Wednesday, October 5, 11:59 pm

1.

Let X, X1,...,Xn be independent, expponentially distributed random variables with parameter λ.

  1. Find the median of X, i.e. a number m such that P(X < m) = 0.5, in terms of λ.
  2. Let the random variable M = median(X1,...,Xn). Usually computation of medians must take into account the possibility of tied values; e.g. see this site. But not here. Why not?
  3. Find P(X1 < 1.2 X2).
  4. Say n = 2, and that Xi is exponentially distributed with parameter λi, i = 1,2. We choose one of the Xi at random, with probability pi. Let T be the result. Find ET and Var(T) in terms of the λi and the pi.

2.

Say the random variable X has an exponential distribution with parameter λ. However, our observation of it is truncated at c, so that our observed value T is min(X,c). Derive a formula for Var(T) as a function of λ and c, using the "Pythagorean Theorem for variance" (Law of Total Variance, SRC p.54). Write simulation code to verify your formula.

3.

The Tower Property says,

E { [E(Y | U,V)] | U} = E(Y | U)

Write simulation code that demonstrates this for a context in which E(Y | U,V) = U + V, so E(Y | U) = U + E(V|U).

A few tips: