Looping With Transition Matrices

Any help with this would be great. I have 2 Matrices, say A and B, which when multiplied make C. I then need to compare the elements of A and C like for like, choosing the largest to make Matrix D. D will then go through the same initial iterations as A until the value of D is unchanged when choosing the largest values.

I have used D=max(A,C) to find D, but I am finding it difficult to loop the process. I have also tried using tf=isequal(A,D), then while tf=0 to run the loop but to no avail.

Any help will be much appreciated.

[AJ von Alt]

Can we see what you have done so far? What part of constructing the loop are you having problems with?

Answer 1

[Amit]

I don't know the process you're doing on D to make it equal to A. However with floating points, there can be carry over like 1.0 and 1.0000001. Here even though 1.0000001 might be an artifact of the process and calculations, but is not equal to 1.0.

You can try

tol = 1e-6;
while (norm(A-D) < tol)

[Kiran]

I am multiplying A, a matrix of probabilities by transition matrix B. Ideally I would need to multiply A by B for a single transition giving another matrix C. I need to compare the probabilities in A and C, choosing the highest in terms of like for like elements ie element A(1,1)=0.2 , C(1,1)=0.4 and so 0.4 is selected.

I need to repeat this process until the probabilities are indifferent.

[Amit]

Try the norm approach and see if it fixes it.