Markov chain exercise

218 Views Asked by At

Hello i have this Markov chain exercise:

Basically we can always move up 1 step, but there is always a possibility that we will go down to the first state 0, the Markov chain consists of N states. When we get to the top we can also stay there with probability p, or go to 0 with probability q(so there are supposed to be 2 p's in the last column)..

The matrix is given by(I have just chosen an N, the N is not fixed but a variable::

\begin{Bmatrix} q & p & 0 & 0 & 0& 0 & 0 & 0 \\ q & 0 & p & 0 & 0& 0 & 0 & 0 \\ q & 0 & 0 & p & 0& 0 & 0 & 0 \\ q & 0 & 0 & 0 & p& 0 & 0 & 0 \\ q & 0 & 0 & 0 & 0& p & 0 & 0 \\ q & 0 & 0 & 0 & 0& 0 & p & 0 \\ q & 0 & 0 & 0 & 0& 0 & 0 & p \\ q & 0 & 0 & 0 & 0& 0 & 0 & p \\ \end{Bmatrix}

I have two questions I am not a able to solve.

1.I am supposed to show that starting in state 0, we will eventually get to state n(I assume that p is bigger than 0). There is a hint to make the last state an absorbing state.

I am not really sure how to solve this. I can maybe solve it with just arguing without any equations like this:? If state N is an absorbing state, then it is also recurrent. The other states are transient. This means that they will only be visited a finite number of times, and hence we have to get to state N at some time. Is this a correct argument?

2. I am supposed to set up equations to that if starting in state i, what is the mean time until I get to state N. And solve these equations, I am able to set them up, but how do I solve them?

Here are the equations: Let $v_i$ be the expected time to reaching state N if we are in state i.

for $i=0,...,N-1$

$v_i=1+q*v_0+p*v_{i+1}$

This equation holds also holds for $v_0$, in order for it to hold for $v_{N-1}$ we note that $v_N=0$.

But I have no idea on how to solve this.

Can someone please help me?

1

There are 1 best solutions below

0
On

(1) The markov chain is irreducible and finite, therefore it is positive recurrent. therefore it will reach state $N$ eventually, when it is started at state $0$.

(2) The equation you have is correct. So,

$v_{n-1} = 1 + qv_0 + pv_n = qv_0 + 1 $

$v_{n-2} = 1 + q v_0 + pv_{n-1} = (p+1)qv_0 + p + 1 $

$v_{n-3} = 1 + qv_0 + pv_{n-2} = (p(p+1)+1)qv_0 + p(p+1)+1$

So, in general $v_{n-i} = f_{i-1}(p)qv_0 + f_{i-1}(p)$ where $f_i(p) = \sum _{j = 0} ^{i} p^{i} $

We also have $(1-q)v_0 = 1 + pv_1 = 1 + p(f_{n-2}(p)qv_0 + f_{n-2}(p)) $. So, we can solve for $v_0$.

There is also an easier way to solve, which you can think about.