TeNPy Forum

Unfortunately, we don't have code for doing this in TeNPy as of now.

There had been some work in that direction, see this topic, but I didn't hear back since that initial zoom meeting.

If you want to implement something in that direction, that would be very welcome

Hi @pasa, from our last discussions, I understood that you had symmetries that can only be broken by 3-site terms, right? In that case even a TEBD-style RandomUnitaryEvolution will be trivial if you fix those symmetries. So yes, to get away from the product states, you need to group sites, run the R...

I'm not sure what you mean with "only interactions of AA BB ABA". The way you define your model, if you consider the sites in a chain as A B A B ... you exclusively get the Heisenberg coupling between A-A sites only, nothing between B -B. Locally on B, you only have a Sz term, so there's n...

For a finite system, this doesn't make sense, so I guess you mean infinite and have some translation invariance by a single site? You can just use truncation parameters with the same chi_min=chi_max , which ultimately fixes the bond dimension exactly. I'd do that only once you have your final MPS, t...

That's right, the grid of the MPO should be np_conserved Arrays. You can also take a look at the mpo_exponential_decay.py example.

The figure 11 is a little bit misleading by stating that this happens between the sweeps, and it seems like we update both left and right environments at once. I chose to simplify in this figure to make it easier to grasp what is going on on the conceptual level - the figure makes clear that you nee...

I got the following, very reasonable question via email: In 1805.00055 there is a picture showing the update of the environment in the iDMRG (figure 11). I understood that between sweeps we are making two copies of the "central" system, and contracting it with the left and right environmen...

Indeed, the MPS is always an open line where a cut of a single virtual leg always divides it into left and right parts - otherwise, the canonical form isometry conditions wouldn't go through e.g. for local expectation values or optimal truncation. When you use infinite MPS, the Model formally has &q...

Sorry for the slow response. I've answered the 165 rather extensively, I hope that clarifies things. If the operator you apply is not exactly unitary, it is to be expected that it changes the norm, isn't it? But you can indeed manually reset psi.norm = 1 by hand if needed. The norm is only relevant ...

In general yes, using Sz or parity conservation can speed up calculations. The actual speedup depends on the number and sizes of symmetry sectors in the tensors - it comes from the fact that the tensors have a block-diagonal form. Parity conservation divides a tensor into 2 subsectors, leading to a ...

A Haar random state has (almost) maximal entanglement and can thus *not* be compressed into an MPS (for large system sizes without significant truncation). If you have a given Haar random state, you could try to convert it to a np_conserved Array and then into an MPS (without charge conservation, so...

The issue lies in different scalings with system size and hence definitions for overlaps that are not what you expect naively. For infinite MPS, the overlap is formally 0, 1, or infinite, and as such not very usefull. Instead, MPS.overlap returns the "overlap per unit cell", to be precise ...

The latter is the recommended way to do a time evolution. You can find the corresponding loop in the source code of the RealTimeEvolution Simulation class in the run_algorithm method.

This is really standard arguments from quantum mechanics: You can use Sz conseration, if [H,S^z] = 0 , where S^z = \sum_i S^z_i . In that case, you *choose* to work in fixed magnetization sectors and write an initial state with S^z |\psi_0\rangle = m \hbar | \psi_0 \rangle , where m\in \mathbb{R} is...

See this thread

Sorry that I couldn't answer earlier, I'd been swamped with other things to finish. You're on the right track, but had some convention issue along the way: First, in TeNPy, the convention is that Jordan Wigner strings extend to the left, not to the right. The MPO you wrote down generates \sum_i \alp...

This is not supported at this point. It's also not quite trivial especially in the case of infinite MPS, since MPOs are often extensive (scaling with L), and multiplying them can give you something quadratic in L.

Thanks for the suggestion, this is indeed the way to go here. See e.g. arXiv:cond-mat/9906019 for details.

On personal laptops, I'd recommend to download anaconda and install the corresponding conda package, then you don't need any build tool yourself.
Note that the latest version 0.9 is by now over a year old, I'll release a new version either today or on Monday....

The XXZ chain by default turns on Sz conservation. Hence, there is no operator 'Sx' or 'Sy', because they don't conserve the Sz charge. Due to Sz conservation, you immediately know that expectation values of Sx and Sy are zero. See also this FAQ question However, note that the correlations <Sx_i Sx_...

The Array class has the methods drop_charge and change_charge . What's missing is functions to do this for each tensor in an MPS (or similarly, an MPO). You could do something along the following snippet: def MPS_drop_charge(psi, charge=None, chinfo=None, permute_p_leg=True): psi_c = psi.copy() psi_...

This should be fixed in the latest github version. I'll release a new version 0.10 (that can be installed with pip/conda) shortly. (Sorry, I didn't update the official releases for over a year, so there have a few fixes accumulated....)

It seems that you use Sz conservation, such that there is no operator 'Sx' or 'Sy', because they don't conserve the Sz charge. Since you model (and hence state) conserve Sz, you immediately know that expectation values of Sx and Sy are zero.
See also this FAQ question

Yes, this looks correct. The line M_tebd = models.model.NearestNeighborModel.from_MPOModel(M) is not really necessary, the Model which you have already is a NearestNeighborModel defining H_bond, so there's no need to create a new model here. In fact, you don't need to define the XXZchainnew here at ...

Indeed, the issue is that the periodic boundary conditions induce a long-range coupling in the model. (The MPS is still open, but I think you already got this.) You don't even need to define a custom model, since there is also the tenpy.models.tf_ising.TFIModel , which is the same as the TFIChain ex...