Markov Chain Monte Carlo question

Click For Summary

Discussion Overview

The discussion revolves around a problem related to Markov Chain Monte Carlo, specifically focusing on finding a regular transition matrix that is not time reversible, meaning it does not satisfy the balance equations. Participants explore the properties of transition matrices and their implications for time reversibility.

Discussion Character

  • Exploratory, Technical explanation, Debate/contested

Main Points Raised

  • One participant seeks help in identifying a regular transition matrix that is not time reversible and suggests that taking the transpose of a transition matrix might be a solution.
  • Another participant clarifies the meaning of time reversibility in the context of transitioning between states in the matrix.
  • A suggestion is made to consider a matrix with linearly dependent rows, which would not be time reversible due to the inability to solve for the inverse.
  • A participant proposes a specific matrix with values .4 and .6, questioning its suitability for the problem.
  • Another participant points out that the determinant of the previously suggested matrix is non-zero, indicating it does not meet the criteria for the problem.
  • A different matrix with repeated rows (.4, .6) is presented as a potential candidate for being non-time reversible.

Areas of Agreement / Disagreement

Participants express differing views on the suitability of specific matrices and the conditions for time reversibility, indicating that the discussion remains unresolved with multiple competing perspectives.

Contextual Notes

Participants do not fully agree on the properties of the matrices discussed, particularly regarding determinants and time reversibility, which may depend on additional assumptions or definitions not fully explored in the thread.

mjt042
Messages
9
Reaction score
0
I was wondering if anyone could help me with this problem dealing with Markov Chain Monte Carlo
-Find a regular transition matrix that is not time reversible, i.e., doesn't satisfy the
balance equations?
My understanding from Markov Chain Monte Carlo is that for the transition matrix to be regular the matrix has to have all positives entries and each row will add up to one. I was thinking the trick to this problem for it not satisfy the balance equation would be to take the transpose of the transition matrix. I was hoping someone could give me a hint if I am on the right track of thinking and where to go from there.

Thanks
 
Physics news on Phys.org
Hey mjt042 and welcome to the forums.

When you mean time reversible do you mean going from transition matrix at state n+1 back to n?
 
  • Like
Likes   Reactions: 1 person
Thanks and yes.
 
Consider a matrix with linearly dependent rows (i.e. a determinant of zero) that still satisfy the probability conditions.

In this situation things are not time reversible since you can not solve for the inverse.
 
  • Like
Likes   Reactions: 1 person
.4 .6
.6 .4 so the matrix to the left would work?
 
No the determinant for this is non-zero.

Consider the matrix

.4 .6
.4 .6
 
Thanks
 

Similar threads

  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 4 ·
Replies
4
Views
2K
  • · Replies 3 ·
Replies
3
Views
3K
  • · Replies 4 ·
Replies
4
Views
5K
Replies
67
Views
8K
  • · Replies 2 ·
Replies
2
Views
2K