Hi Jessica,
It is easy to specify a graph where one bond has a different type, but then you lose translational invariance. Is that a problem for you?
Matthias
On 20 Jan 2010, at 13:42, jessica.alfonsi@unipd.it wrote:
Hi everyone, I have posted some days ago about full diagonalization with antiperiodic boundary conditions for a Hubbard chain but I havent't received no answer yet. Maybe I was not much clear with my question. The point with my question is that I'd like to specify antiperiodic boundary conditions for some specific sites (i.e. a bond term t#*fermion_hop(1,N), where sites 1 and N are the chain ends) in the input model xml file. This means that the terms related to boundary conditions have to be specified separately from the usual bond terms with fermionic hopping. Is it possible to do that in the xml model file ? How can I do that in a C++ code compiled against ALPS libraries ? This would be very important for specifying lattices and models with the highest flexibility, since later I'd like to write a matrix free ED code for my 2D custom lattice... Could you give me some insight please ? Best regards
Jessica Alfonsi
-- Dr.ssa Jessica Alfonsi Dipartimento di Scienze Chimiche Università di Padova via Marzolo 1 35100 Padova (Italy) +39 0498275726 e-mail jessica.alfonsi@unipd.it