Group discussion

[Eduard Vakarin]

Hello everybody,

I am pleased to participate in the discussion.

First of all, I would like to comment on the arguments of Prof. Abe:
"It seems to me that the main issue at the moment is concerned with
generalized entropies in connection with non-exponential distributions."

The main question (on my opinion) is why should we change the entropy
measure in order to get a given distribution in the easiest way (e.g. with
a linear constraint)? A non-exponential distribution corresponding to a
stationary non-equilibrium state can be obtained from the Shannon entropy
with the constraints "encoding" the non-equilibrium situation. As I have
already mentioned I have tried to demonstrate that in Phys. Rev. E 74,
036120
(2006). On my opinion this approach is quite natural because the
constraints carry the physical information on the system of interest. On
the other
hand, generalizing the entropy we are trying to hide this physics in a
new functional.

I am preparing my notes on a relation of the maxent and the varentropy
proposed by Alexandre Wang. I hope to finish soon and the I will give
an arXiv reference.

Best regards to everybody

Eduard Vakarin