MATH 180B Homework 7
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
MATH 180B Homework 7
2022
1. (Wright-Fisher model) Consider a model of a population with constant size N and types E = (1, 2}. Let Xn be the number of type 1 individuals at time n. The evolution is described as follows: the distribution of Xn+0 is
Binomial(N, ) conditioned on the event Xn = k , k e (0, 1, . . . , N }. l)( Write down the transition matrix of the Markov chain Xo , X0 , ... l))( Is this Markov chain irreducible? Does it have stationary distributions?
2. (A variation of Wright-Fisher model) Similar to Problem 1, consider a model of a population with constant size N and types E = (1, 2}. Let Xn be the number of type 1 individuals at time n. The evolution is now changed to: there is a parameter u e (0, 1), the distribution of Xn+0 is Binomial iN, u + (1 _ u) i1 _ 、、conditioned on the event Xn = k , k e (0, 1, . . . , N }. Observe that 0, N are not absorbing states for this chain.
l)( Show that the Markov chain (Xn ) converges to its unique stationary dis- tribution π. (Hint: Apply the convergence theorem, you do not need to
compute the stationary distribution.)
l))( Compute limn → ~ 匝x [Xn] for x e (0, 1, ..., N }.
3. Let (Xn ) be the Markov chain on the state space (0, 1, 2, ...} with transi- tion probability
P (0, 0) = 1/2, P (0, 1) = 1/2,
P (i, i + 1) = /1 _ 、 , P (i, i _ 1) = /1 + 、 , i e (1, 2...}.
l)( Determine whether this Markov chain has stationary distribution. If your answer is yes, give an expression for the stationary distribution; if your answer is no, give reasons why this Markov chain doesn’t admit any sta- tionary distribution.
l))( Based on part (i), determine whether the states of this Markov chain are transient, positive recurrent or null recurrent.
4. Suppose in a branching process, the offspring distribution is a shifted geometric distribution with probability mass function
pk = p(1 _ p)k , k = 0, 1, 2, ....
Compute the probability that starting from one individual that the chain gets absorbed at 0.
5. Let Z be the total family size in a branching process whose offspring distri- bution has mean µ < 1. Given that Xo = 1, show that
匝o (Z) = 1
2022-03-06