It can happen that two equivariant sets with atoms are related by a finitely supported bijection, but not by an equivariant bijection.
It is easy to see that, for equivariant sets with equality atoms and
, a finitely supported function
may exist even if an equivariant function from
to
does not. For example, let
be a singleton and
the set of atoms
; one can map the unique element to an atom, but one needs to choose the atom first.
It is less clear what happens if the finitely supported function is a bijection: does an equivariant bijection exist as well?
The answer is negative. For example, consider
,
choose an atom and define a function
by:
for
.
It is easy to see that is a bijection supported by
. However, no equivariant bijection (or even an equivariant function) from
to
exists, since
contains an element with empty support and
does not.
Interestingly though, if and
are orbit-finite, then finitely supported bijections can be smoothed into equivariant ones. This was proved by Joanna, who in the near future will have written a post about it.