issp-center-dev / TeNeS

Massively parallel tensor network solver
http://www.pasums.issp.u-tokyo.ac.jp/tenes/en
GNU General Public License v3.0
46 stars 11 forks source link

How to calculate the norm per site of an iTPS ? #95

Closed sunzhaoyu2000 closed 3 weeks ago

sunzhaoyu2000 commented 3 weeks ago

Hi everyone,

I have a question about the norm of an iTPS. I try to consider a 2D square lattice with size $M\times N$, where the quantum state is described by a infinite-size translation-invariant iTPS $\vert \Psi \rangle$.

To make my question clear, I have found a figure in the "TeNeS-master/docs/sphinx/img" directory and I have attached it as follows.

g944


It is expected that the norm can be expressed as $$\langle \Psi \vert \Psi \rangle = k^{M\times N},$$ where $k$ is called "the norm per site".

My question is how to figure out this coefficient $k$ ?

In file "01_transverse_field_ising/output/onesite_obs.dat", it reads

The meaning of each column is the following: $1: op_group $2: site_index $3: real $4: imag The names of op_group are the following: 0: Sz
1: Sx
-1: norm 0 0 1.43822572220935002e-01 0.00000000000000000e+00 0 1 1.43822567502205789e-01 0.00000000000000000e+00 0 2 1.43822577490189368e-01 0.00000000000000000e+00 0 3 1.43822572827914413e-01 0.00000000000000000e+00 1 0 4.51957838213799423e-01 0.00000000000000000e+00 1 1 4.51957837144178809e-01 0.00000000000000000e+00 1 2 4.51957758107002339e-01 0.00000000000000000e+00 1 3 4.51957756949986855e-01 0.00000000000000000e+00 -1 0 8.02176941013978739e-01 0.00000000000000000e+00 -1 1 8.02176538960576391e-01 0.00000000000000000e+00 -1 2 8.02178199879971077e-01 0.00000000000000000e+00 -1 3 8.02177795112917202e-01 0.00000000000000000e+00

For the last four lines, the number "8.02e-01" seems to be obtained by simple contracting the corner transfer matrices and the edge tensors in Fig. (c), and its role seems to be the norm of the reduced density matrices for the local $2\times 2$ sub-squares.

Thus, my question is that, is it possible to figure out the coefficient $k$ ? I guess $k$ is still closely related to Fig. (c) in some way, but I am not sure about the details.

Any suggestion or comment would be greatly appreciated.

sunzhaoyu2000 commented 3 weeks ago

Hi everyone,

The above problem has been solved in this article perfectly :

PHYSICAL REVIEW B 97, 115161 (2018) Infinite projected entangled-pair state algorithm for ruby and triangle-honeycomb lattices Saeed S. Jahromi et al. DOI: 10.1103/PhysRevB.97.115161

Thank you.