A frog is sitting on a lily pad in a lake.
Every second, the frog jumps either left or right 1 lily pad with equal probability.
There is only 1 lily pad to its left, but there is an infinite amount of lily pads to its right.
What is the expected value of the number of seconds before it reaches the leftmost lily pad?
Enter -1 if you think the expected value is infinite.
This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try
refreshing the page, (b) enabling javascript if it is disabled on your browser and,
finally, (c)
loading the
non-javascript version of this page
. We're sorry about the hassle.
Can you explain a bit more?
Log in to reply
Not easily. There is a massive amount of book work about Markov chains here, and its application to random walks, and it would take pages to type it all out.
There is plenty of stuff out there on the web. Try hunting for the "symmetric one-dimensional random walk", which is null recurrent, as opposed to the asymmetric one, which is transient. There are also higher dimensional random walks.
Log in to reply
and I thought random walks are easy :/
Log in to reply
@Kunal Gupta – I can give you an "easy" explanation of the problem, if not of the solution I posted. Let E n be the expected number of moves to reach the left-most lilypad, given that the frog starts n lilypads to the right. Of course E 0 = 0 . By conditioning on the outcome of the first jump, we have that E n = 2 1 ( 1 + E n − 1 ) + 2 1 ( 1 + E n + 1 ) = 1 + 2 1 ( E n − 1 + E n + 1 ) n ≥ 1 Let us assume that E 1 is finite (this is the number we are asked to calculate). It is then clear by induction that E n is finite for all n ≥ 0 (just use the recurrence relation to calculate E 2 , E 3 , and so on up).
If we solve the recurrence relation, we must have E n = A + B n − n 2 for some constants A , B . Since E 0 = 0 , we deduce that A = 0 , and hence E n = B n − n 2 . Even though we do not know what B is, it is clear that that formula must have E n negative for large enough n , which is not possible.
Thus we deduce that E 1 is infinite.
Problem Loading...
Note Loading...
Set Loading...
The one-dimensional symmetric random walk is null recurrent, and therefore all expected hitting times are infinite.