Author 
Message 
Chip Eastham science forum Guru
Joined: 01 May 2005
Posts: 412

Posted: Tue May 30, 2006 3:29 pm Post subject:
Re: 3 Dimensional Random Walk



Jim Dars wrote:
Quote:  "Barry Schwarz" <schwarzb@doezl.net> wrote in message
news:ssoh72tir6bict6kvvqmhgqgpldp7mqo6r@4ax.com...
On Sat, 27 May 2006 13:32:50 0700, "Jim Dars" <jimdars@comcast.net
wrote:
Hi All,
Feller gives in Volume I, "An Introduction to Probability Theory and It's
Applications", that if one starts at the origin and moves at random the
probability of returning to the origin in one or two dimensions is one
and
in three dimensions it is about .35. (In two dimensions there are 4
random
choices and in 3 dimensions, 6.)
Question: In 3 dimensions if you start at the point (10000,10000,10000)
do
you arrive at a different answer?
Does your question refer to returning to the starting point or
returning to the origin?
Remove del for email
Hi Barry,
It concerns "returning" (getting there the first time) to the origin.
Actually, I suspect that, in the three dimensional case, the starting point
influences the probability of "returning" to the origin. I'm wondering what
the Group thinks (knows).
Best wishes, Jim

I can recommend this (free) monograph by Doyle and Snell on the
subject:
http://front.math.ucdavis.edu/math.PR/0001057
In two dimensions, a random walk on the "graph" connecting
the usual integer lattice points visits every point in the plane
with probability 1. So in that setting it doesn't matter whether
we ask about "returning" to the origin or going from one point
to another, the probability is the same. This can be deduced
from the divergence of SUM 1/n.
In three dimensions, as Jim quotes Feller, the chance of
returning to an original starting point is positive and strictly
less than 1 (about .35). This is related to the convergence
of SUM 1/(n^2), although the connection is not elegant when
made rigorous. The search for a simple connection motivates
much of the monograph above.
In short we can define a function that gives the probability
that a random walk starting at the origin will visit (at some
subsequent step) any specific point in the integer lattice.
Evaluation of this function at (1000,1000,1000) will give by
symmetry the answer to Jim's question, and in general
this function is monotone decreasing as we go farther
from the origin.
regards, chip 

Back to top 


Jim Dars science forum beginner
Joined: 03 Jul 2005
Posts: 41

Posted: Wed May 31, 2006 9:03 pm Post subject:
Re: 3 Dimensional Random Walk



"Chip Eastham" <hardmath@gmail.com> wrote in message
news:1149002994.240544.237280@38g2000cwa.googlegroups.com...
Quote: 
Jim Dars wrote:
"Barry Schwarz" <schwarzb@doezl.net> wrote in message
news:ssoh72tir6bict6kvvqmhgqgpldp7mqo6r@4ax.com...
On Sat, 27 May 2006 13:32:50 0700, "Jim Dars" <jimdars@comcast.net
wrote:
Hi All,
Feller gives in Volume I, "An Introduction to Probability Theory and
It's
Applications", that if one starts at the origin and moves at random
the
probability of returning to the origin in one or two dimensions is
one
and
in three dimensions it is about .35. (In two dimensions there are 4
random
choices and in 3 dimensions, 6.)
Question: In 3 dimensions if you start at the point
(10000,10000,10000)
do
you arrive at a different answer?
Does your question refer to returning to the starting point or
returning to the origin?
Remove del for email
Hi Barry,
It concerns "returning" (getting there the first time) to the origin.
Actually, I suspect that, in the three dimensional case, the starting
point
influences the probability of "returning" to the origin. I'm wondering
what
the Group thinks (knows).
Best wishes, Jim
I can recommend this (free) monograph by Doyle and Snell on the
subject:
http://front.math.ucdavis.edu/math.PR/0001057
In two dimensions, a random walk on the "graph" connecting
the usual integer lattice points visits every point in the plane
with probability 1. So in that setting it doesn't matter whether
we ask about "returning" to the origin or going from one point
to another, the probability is the same. This can be deduced
from the divergence of SUM 1/n.
In three dimensions, as Jim quotes Feller, the chance of
returning to an original starting point is positive and strictly
less than 1 (about .35). This is related to the convergence
of SUM 1/(n^2), although the connection is not elegant when
made rigorous. The search for a simple connection motivates
much of the monograph above.
In short we can define a function that gives the probability
that a random walk starting at the origin will visit (at some
subsequent step) any specific point in the integer lattice.
Evaluation of this function at (1000,1000,1000) will give by
symmetry the answer to Jim's question, and in general
this function is monotone decreasing as we go farther
from the origin.
regards, chip
Hi Chip, 
Thank you. I suspected as much in that if you start at the origin you have
a 1/6 probability of reaching the origin on the second move. It's
interesting to learn that, as you move the starting point away from the
origin, the probability of reaching the origin diminishes. A little
reflection does indicate this as a logical conclusion, but seemingly logical
conclusions often blow up in your face.
Best wishes, Jim 

Back to top 


Mike Amling science forum Guru
Joined: 05 May 2005
Posts: 525

Posted: Thu Jun 08, 2006 10:50 pm Post subject:
Re: 3 Dimensional Random Walk



In article <1149002994.240544.237280@38g2000cwa.googlegroups.com>,
Chip Eastham <hardmath@gmail.com> wrote:
Quote: 
In short we can define a function that gives the probability
that a random walk starting at the origin will visit (at some
subsequent step) any specific point in the integer lattice.
Evaluation of this function at (1000,1000,1000) will give by
symmetry the answer to Jim's question, and in general
this function is monotone decreasing as we go farther
from the origin.

THe probability that a random walk starting at the origin will
eventually return to the origin should be the same as the probability
that a random walk starting at (1,0,0), (0,1,0), (0,0,1), (1,0,0),
(0,1,0), or (0,0,1) will eventually reach the origin.

Please reply to:  "Any sufficiently advanced incompetence is
pciszek at panix dot com  indistinguishable from malice."
Autoreply is disabled  

Back to top 


Google


Back to top 



The time now is Fri Oct 20, 2017 2:21 pm  All times are GMT

