TDVPEngine

Inheritance Diagram

Inheritance diagram of tenpy.algorithms.tdvp.TDVPEngine

Methods

TDVPEngine.__init__(psi, model, options[, ...])

TDVPEngine.get_resume_data([...])

Return necessary data to resume a run() interrupted at a checkpoint.

TDVPEngine.resume_run()

Resume a run that was interrupted.

TDVPEngine.run()

(Real-)time evolution with TDVP.

TDVPEngine.run_one_site([N_steps])

Run the TDVP algorithm with the one site algorithm.

TDVPEngine.run_two_sites([N_steps])

Run the TDVP algorithm with two sites update.

TDVPEngine.set_anonymous_svd(U, new_label)

Relabel the svd.

TDVPEngine.sweep_left_right()

Performs the sweep left->right of the second order TDVP scheme with one site update.

TDVPEngine.sweep_left_right_two()

Performs the sweep left->right of the second order TDVP scheme with two sites update.

TDVPEngine.sweep_right_left()

Performs the sweep right->left of the second order TDVP scheme with one site update.

TDVPEngine.sweep_right_left_two()

Performs the sweep left->right of the second order TDVP scheme with two sites update.

TDVPEngine.theta_svd_left_right(theta)

Performs the SVD from left to right.

TDVPEngine.theta_svd_right_left(theta)

Performs the SVD from right to left.

TDVPEngine.update_s_h0(s, H, dt)

Update with the zero site Hamiltonian (update of the singular value)

TDVPEngine.update_theta_h1(Lp, Rp, theta, W, dt)

Update with the one site Hamiltonian.

TDVPEngine.update_theta_h2(Lp, Rp, theta, ...)

Update with the two sites Hamiltonian.

Class Attributes and Properties

TDVPEngine.TDVP_params

TDVPEngine.time_dependent_H

whether the algorithm supports time-dependent H

TDVPEngine.verbose

class tenpy.algorithms.tdvp.TDVPEngine(psi, model, options, environment=None, **kwargs)[source]

Bases: tenpy.algorithms.algorithm.TimeEvolutionAlgorithm

Time dependent variational principle algorithm for MPS.

Deprecated since version 0.6.0: Renamed parameter/attribute TDVP_params to options.

Parameters
  • psi – Same as for Algorithm.

  • model – Same as for Algorithm.

  • options – Same as for Algorithm.

  • **kwargs – Same as for Algorithm.

  • environment – Initial environment. If None (default), it will be calculated at the beginning.

Options

config TDVP
option summary

active_sites

The number of active sites to be used for the time evolution. [...]

dt (from TimeEvolutionAlgorithm) in TimeEvolutionAlgorithm

Minimal time step by which to evolve.

lanczos_options

Lanczos options as described in :cfg:config:`Lanczos`.

N_steps (from TimeEvolutionAlgorithm) in TimeEvolutionAlgorithm

Number of time steps `dt` to evolve by in :meth:`run`. [...]

start_time (from TimeEvolutionAlgorithm) in TimeEvolutionAlgorithm

Initial value for :attr:`evolved_time`.

trunc_params

Truncation parameters as described in :func:`~tenpy.algorithms.truncation.t [...]

option active_sites

The number of active sites to be used for the time evolution. If set to 1, run_one_site() is used. The bond dimension will not increase! If set to 2, run_two_sites() is used.

option trunc_params: dict

Truncation parameters as described in truncate()

option lanczos_options: dict

Lanczos options as described in Lanczos.

options

Optional parameters.

Type

dict

evolved_time

Indicating how long psi has been evolved, psi = exp(-i * evolved_time * H) psi(t=0).

Type

float | complex

psi

The MPS, time evolved in-place.

Type

MPS

environment

The environment, storing the LP and RP to avoid recalculations.

Type

MPOEnvironment

lanczos_options

Options passed on to LanczosEvolution.

Type

Config

run()[source]

(Real-)time evolution with TDVP.

run_one_site(N_steps=None)[source]

Run the TDVP algorithm with the one site algorithm.

Warning

Be aware that the bond dimension will not increase!

Parameters

N_steps (integer. Number of steps) –

run_two_sites(N_steps=None)[source]

Run the TDVP algorithm with two sites update.

The bond dimension will increase. Truncation happens at every step of the sweep, according to the parameters set in trunc_params.

Parameters

N_steps (integer. Number of steps) –

sweep_left_right()[source]

Performs the sweep left->right of the second order TDVP scheme with one site update.

Evolve from 0.5*dt.

sweep_left_right_two()[source]

Performs the sweep left->right of the second order TDVP scheme with two sites update.

Evolve from 0.5*dt

sweep_right_left()[source]

Performs the sweep right->left of the second order TDVP scheme with one site update.

Evolve from 0.5*dt

sweep_right_left_two()[source]

Performs the sweep left->right of the second order TDVP scheme with two sites update.

Evolve from 0.5*dt

update_theta_h1(Lp, Rp, theta, W, dt)[source]

Update with the one site Hamiltonian.

Parameters
  • Lp (Array) – tensor representing the left environment

  • Rp (Array) – tensor representing the right environment

  • theta (Array) – the theta tensor which needs to be updated

  • W (Array) – MPO which is applied to the ‘p’ leg of theta

update_theta_h2(Lp, Rp, theta, W0, W1, dt)[source]

Update with the two sites Hamiltonian.

Parameters
theta_svd_left_right(theta)[source]

Performs the SVD from left to right.

Parameters

theta (tenpy.linalg.np_conserved.Array) – the theta tensor on which the SVD is applied

set_anonymous_svd(U, new_label)[source]

Relabel the svd.

Parameters

U (tenpy.linalg.np_conserved.Array) – the tensor which lacks a leg_label

theta_svd_right_left(theta)[source]

Performs the SVD from right to left.

Parameters

theta (tenpy.linalg.np_conserved.Array,) – The theta tensor on which the SVD is applied

update_s_h0(s, H, dt)[source]

Update with the zero site Hamiltonian (update of the singular value)

Parameters
  • s (tenpy.linalg.np_conserved.Array) – representing the singular value matrix which is updated

  • H (H0_mixed) – zero site Hamiltonian that we need to apply on the singular value matrix

  • dt (complex number) – time step of the evolution

get_resume_data(sequential_simulations=False)[source]

Return necessary data to resume a run() interrupted at a checkpoint.

At a checkpoint, you can save psi, model and options along with the data returned by this function. When the simulation aborts, you can resume it using this saved data with:

eng = AlgorithmClass(psi, model, options, resume_data=resume_data)
eng.resume_run()

An algorithm which doesn’t support this should override resume_run to raise an Error.

Parameters

sequential_simulations (bool) – If True, return only the data for re-initializing a sequential simulation run, where we “adiabatically” follow the evolution of a ground state (for variational algorithms), or do series of quenches (for time evolution algorithms); see run_seq_simulations().

Returns

resume_data – Dictionary with necessary data (apart from copies of psi, model, options) that allows to continue the simulation from where we are now. It might contain an explicit copy of psi.

Return type

dict

resume_run()[source]

Resume a run that was interrupted.

In case we saved an intermediate result at a checkpoint, this function allows to resume the run() of the algorithm (after re-initialization with the resume_data). Since most algorithms just have a while loop with break conditions, the default behaviour implemented here is to just call run().