peropq.variational_unitary¶

Classes¶

class representing the variational unitary ansatz.

class peropq.variational_unitary.VariationalUnitary(hamiltonian, number_of_layer, time)¶

class representing the variational unitary ansatz.

Parameters:

pauli_string_list: collections.abc.Sequence[peropq.pauli_bitstring.PauliString]¶

update_theta(new_array)¶

Update theta ensuring that the condition Sum_i theta_i dt_i= is ensured.

Parameters:: new_array (numpy.typing.NDArray) – the new array containing the variational parameters. It’s shape must be (R - 1, n_terms).
Return type:: None

get_initial_trotter_vector()¶

Get the variational parameters corresponding to the Trotterization. Useful to initialize the optimization.

flatten_theta(theta)¶

Returns the variational parameters as flatten (R-1)*n_terms array. Useful to pass to a minimization function.

unflatten_theta(flat_theta)¶

Returns the flattened variational parameters as an array with shape (depth,n_terms).

set_theta_to_trotter()¶

Sets the variational parameters to the Trotter parameters.

chi_tensor(left_indices, right_indices)¶

Vectorized function to calculate all the chi coefficient at once.

Parameters:

left_indices (numpy.typing.NDArray) – indices of the left tensor which give non-zero contributions in the calculation of chi.
right_indices (numpy.typing.NDArray) – indices of the right tensor which give non-zero contributions in the calculation of chi.

Returns:

numpy array representing the tensor used in the norm calculation.

Return type:

numpy.typing.NDArray

calculate_traces()¶

Calculate the trace tensor.

c2_squared(theta=())¶

Perturbative 2-norm.

Parameters:: theta (numpy.typing.ArrayLike) – parameters of the variational unitary.
Returns:: the perturbative approximation of the 2-norm difference between the exact and the variational representation.
Return type:: float