- #1
fluidistic
Gold Member
- 3,924
- 261
Homework Statement
Hi guys, I'm absolutely desperate on the following problem:
Consider a random walker who can make steps only to neighbor sites in "D" dimensions where D is an arbitrary natural number. Assume that the distance between 2 adjacent sites is the same for all sites and that the probabilty for the walker to choose between any adjacent site is the same. Let [itex]\vec r[/itex] be a particular site and N be the number of steps the walker has done.
What is the probability to find the random walker at [itex]\vec r[/itex] after N steps if he starts initially at the origin?
Answer to this question using:
1)the assumption that the walk can be described as a Markov chain in D dimensions.
2)the distribution function of the displacement of the walker after N independent steps, [itex]\vec r = \vec r _1 + ... + \vec r _N[/itex].
Homework Equations
Probably a lot! Not really sure.
The Attempt at a Solution
I tried part 1) so far but did not go far at all.
I know that the probability to choose any adjacent site is [itex]\frac{1}{2D}[/itex] where D is the dimension of the space the walker walks in.
So they are asking me [itex]P_N(\vec r)[/itex]. I'll use the notation [itex]\vec r=r_1 \hat x_1 +... + r_D \hat x_D[/itex]; that's the position of the walker after N steps.
I have the initial condition that [itex]P_0(\vec r )=\vec 0[/itex]. Since I'm given an initial condition on [itex]P_N(\vec r )[/itex] and that I'm asked to find [itex]P_N (\vec r )[/itex] I can "smell" that I'll have to solve a differential equation or something like that, but I've no idea how to find it.
Now I know that for a Markovian process, the probability that the walker will go to say [itex]P _N(\vec r)[/itex] does not depend on its past but only on its present. Namely only on the "state" [itex]P_ {N-1} (\vec l )[/itex] where [itex]\vec l[/itex] is the site of the walker prior to [itex]\vec r[/itex].
If I'm not wrong, then, I guess [itex]P_N (\vec r)=\frac{1}{2D} P_{N-1} (\vec l )[/itex].
But then if I take [itex]P_{N-1} (\vec l )[/itex] it will depend only on the previous state. In the end I get the wrong result that [itex]P_N ( \vec r ) = \left ( \frac{1}{2D} \right ) ^{N} P_0 (0)=\left ( \frac{1}{2D} \right ) ^{N}[/itex].
I know this result is wrong but I don't know what I'm doing wrong. Any help will be appreciated.