Engine

• full name: tenpy.algorithms.tdvp.Engine

• parent module: tenpy.algorithms.tdvp

• type: class

class tenpy.algorithms.tdvp.Engine(psi, model, TDVP_params, environment=None)

Bases: object

Time dependant variational principle ‘Engine’.

You can call run_one_site() for single-site TDVP, or run_two_sites() for two-site TDVP.

Parameters

psiMPS

Initial state to be time evolved. Modified in place.

modelMPOModel

The model representing the Hamiltonian for which we want to find the ground state.

TDVP_paramsdict

Further optional parameters as described in the following table. Use verbose>0 to print the used parameters during runtime.

key	type	description
start_time	float	Initial value for evolved_time
dt	float	Time step of the Trotter error
trunc_params	dict	Truncation parameters as described in truncate()

environment:class:’~tenpy.networks.mpo.MPOEnvironment` | None

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

Attributes

verboseint

Level of verbosity (i.e. how much status information to print); higher=more output.

evolved_timefloat | complex

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

psiMPS

The MPS, time evolved in-place.

TDVP_params: dict

Optional parameters, see run() and run_GS() for more details.

environmentMPOEnvironment

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

Methods

run_one_site(self[, N_steps])	Run the TDVP algorithm with the one site algorithm.
run_two_sites(self[, N_steps])	Run the TDVP algorithm with two sites update.
set_anonymous_svd(self, U, new_label)	Relabel the svd.
sweep_left_right(self)	Performs the sweep left->right of the second order TDVP scheme with one site update.
sweep_left_right_two(self)	Performs the sweep left->right of the second order TDVP scheme with two sites update.
sweep_right_left(self)	Performs the sweep right->left of the second order TDVP scheme with one site update.
sweep_right_left_two(self)	Performs the sweep left->right of the second order TDVP scheme with two sites update.
theta_svd_left_right(self, theta)	Performs the SVD from left to right.
theta_svd_right_left(self, theta)	Performs the SVD from right to left.
update_s_h0(self, s, H, dt)	Update with the zero site Hamiltonian (update of the singular value)
update_theta_h1(self, Lp, Rp, theta, W, dt)	Update with the one site Hamiltonian.
update_theta_h2(self, Lp, Rp, theta, W0, W1, dt)	Update with the two sites Hamiltonian.

run_one_site(self, N_steps=None)

Run the TDVP algorithm with the one site algorithm.

Warning

Be aware that the bond dimension will not increase!

Parameters

N_stepsinteger. Number of steps

run_two_sites(self, N_steps=None)

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_stepsinteger. Number of steps

sweep_left_right(self)

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

Evolve from 0.5*dt.

sweep_left_right_two(self)

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

Evolve from 0.5*dt

sweep_right_left(self)

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

Evolve from 0.5*dt

sweep_right_left_two(self)

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

Evolve from 0.5*dt

update_theta_h1(self, Lp, Rp, theta, W, dt)

Update with the one site Hamiltonian.

Parameters

LpArray

tensor representing the left environment

RpArray

tensor representing the right environment

thetaArray

the theta tensor which needs to be updated

WArray

MPO which is applied to the ‘p’ leg of theta

update_theta_h2(self, Lp, Rp, theta, W0, W1, dt)

Update with the two sites Hamiltonian.

Parameters

Lptenpy.linalg.np_conserved.Array

tensor representing the left environment

Rptenpy.linalg.np_conserved.Array

tensor representing the right environment

thetatenpy.linalg.np_conserved.Array

the theta tensor which needs to be updated

Wtenpy.linalg.np_conserved.Array

MPO which is applied to the ‘p0’ leg of theta

W1tenpy.linalg.np_conserved.Array

MPO which is applied to the ‘p1’ leg of theta

theta_svd_left_right(self, theta)

Performs the SVD from left to right.

Parameters

theta: :class:`tenpy.linalg.np_conserved.Array`

the theta tensor on which the SVD is applied

set_anonymous_svd(self, U, new_label)

Relabel the svd.

Parameters

Utenpy.linalg.np_conserved.Array

the tensor which lacks a leg_label

theta_svd_right_left(self, theta)

Performs the SVD from right to left.

Parameters

thetatenpy.linalg.np_conserved.Array,

The theta tensor on which the SVD is applied

update_s_h0(self, s, H, dt)

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

Parameters

stenpy.linalg.np_conserved.Array

representing the singular value matrix which is updated

HH0_mixed

zero site Hamiltonian that we need to apply on the singular value matrix

dtcomplex number

time step of the evolution