How to Compute the Sum in the Stationary Distribution of a Markov Chain?

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
1 reply · 2K views
TaPaKaH
Messages
51
Reaction score
0
Suppose we have a Markov chain with stationary distributions ##p_n=\frac{a}{nb+c}p_{n-1}## for ##n\in\mathbb{N}## where ##a,b## and ##c## are some positive constants.
It follows that ##p_n=p_0\prod_{i=1}^n\frac{a}{ib+c}##. Normalisation yields ##1=p_0\sum_{n=0}^\infty\prod_{i=1}^n\frac{a}{ib+c}## so ##p_0=\left(\sum_{n=0}^\infty\prod_{i=1}^n\frac{a}{ib+c}\right)^{-1}##.

Question: how can one compute the sum in the brackets?
 
Last edited:
Physics news on Phys.org
TaPaKaH said:
Suppose we have a Markov chain with stationary distributions ##p_n=\frac{a}{nb+c}p_{n-1}## for ##n\in\mathbb{N}## where ##a,b## and ##c## are some positive constants.
It follows that ##p_n=p_0\prod_{i=1}^n\frac{a}{ib+c}##. Normalisation yields ##1=p_0\sum_{n=0}^\infty\prod_{i=1}^n\frac{a}{ib+c}## so ##p_0=\left(\sum_{n=0}^\infty\prod_{i=1}^n\frac{a}{ib+c}\right)^{-1}##.

Question: how can one compute the sum in the brackets?
Well, let's think about this. What can we do to simplify $$\prod_{1\leq i\leq n}\frac{a}{bi+c}$$ into "workable" terms? Can you give us an attempt?