variables casts the operator in a form compatible with the theorem, thus
establishing the impossibility of phase transitions in this model (and in fact
in a much wider class). Interestingly, the same problem arises in van Hove’s
theorem;
(8)
van Hove writes first his general transfer operator in a non-
compact form, but he is able to rewrite it as a compact operator and to
prove his theorem. The difference with respect to the Chui–Weeks’s model
is that in this case it is possible to calculate the spectrum and prove that
there is actually an eigenvalue crossing. These considerations indicate that
non-compactness of the transfer operator merely excludes it from the
theorem, but is not enough to say anything definite about the possibility of
phase transitions.
Do'stlaringiz bilan baham: