Welcome to MathOverflow

Question

-2

1

It is possible to define the relative entropy between to diffusion processes using Girsanov theorem only in the case where the diffusion coefficients are the same. Could you explain why it is not posssible to define a relative entropy between the laws of the two following processes: $$dx_1(t)=f(x_1(t))dt +\sigma_1 dW_t$$ and $$dx_2(t)=f(x_2(t))dt +\sigma_2 dW_t$$ when $\sigma_1 \neq \sigma_2$ ?

Beyond the problem of absolute continuity, is not there a way to compare both laws ?

flag	asked Nov 6 2011 at 18:33 unknown (google) 249●5

1

You noticed that the Wiener measures were singulars I think then that there is nothing more to say. What is entropy between singular measures ? – The Bridge Nov 6 2011 at 22:09

This is precisely the point ! Is it really the end of the story? or the begining of a new one? – unknown (google) Nov 6 2011 at 22:35

brownck · Answer 1 · 2012-08-17 04:43:13Z UTC

-1

This is not an answer to your question, but can you explain and perhaps give a reference to your statement:

It is possible to define the relative entropy between to diffusion processes using Girsanov theorem only in the case where the diffusion coefficients are the same.

answered Aug 17 at 4:43

brownck
1

Ask this as a separate question. – Didier Piau Aug 17 at 8:00

Relative entropy for diffusion processes

Remember to vote up questions/answers you find interesting or helpful (requires 15 reputation points)

1 Answer

You can accept an answer to one of your own questions by clicking the check mark next to it. This awards 15 reputation points to the person who answered and 2 reputation points to you.