qml.grouping

This subpackage defines functions and classes for Pauli-word partitioning functionality used in measurement optimization.

A Pauli word is defined as \(P_J = \prod_{i=1}^{N}\sigma_i^{(J)}\), where \(\sigma_i^{(J)}\) is one of the Pauli operators (PauliX, PauliY, PauliZ) or identity (Identity) acting on the \(i^{th}\) qubit.

Pauli words can be used for expressing a qubit Hamiltonian. A qubit Hamiltonian has the form \(H_{q} = \sum_{J} C_J P_J\) where \(C_{J}\) are numerical coefficients, and \(P_J\) are Pauli words.

A list of Pauli words can be partitioned according to certain grouping strategies. As an example, the group_observables() function partitions a list of observables (Pauli operations and tensor products thereof) into groupings according to a binary relation (e.g., qubit-wise commuting):

>>> observables = [qml.PauliY(0), qml.PauliX(0) @ qml.PauliX(1), qml.PauliZ(1)]
>>> obs_groupings = group_observables(observables)
>>> obs_groupings
[[Tensor(PauliX(wires=[0]), PauliX(wires=[1]))],
 [PauliY(wires=[0]), PauliZ(wires=[1])]]

The \(C_{J}\) coefficients for each \(P_J\) Pauli word making up a Hamiltonian can also be specified along with further options, such as the Pauli-word grouping method (e.g., qubit-wise commuting) and the underlying graph-colouring algorithm (e.g., recursive largest first) used for creating the groups of observables:

>>> obs = [qml.PauliY(0), qml.PauliX(0) @ qml.PauliX(1), qml.PauliZ(1)]
>>> coeffs = [1.43, 4.21, 0.97]
>>> obs_groupings, coeffs_groupings = group_observables(obs, coeffs, 'qwc', 'rlf')
>>> obs_groupings
[[Tensor(PauliX(wires=[0]), PauliX(wires=[1]))],
 [PauliY(wires=[0]), PauliZ(wires=[1])]]
>>> coeffs_groupings
[[4.21], [1.43, 0.97]]

pennylane.grouping Package

This subpackage defines functions and classes for Pauli-word partitioning functionality used in measurement optimization.

Functions

are_identical_pauli_words(pauli_1, pauli_2)	Performs a check if two Pauli words have the same wires and name attributes.
binary_to_pauli(binary_vector[, wire_map])	Converts a binary vector of even dimension to an Observable instance.
diagonalize_pauli_word(pauli_word)	Transforms the Pauli word to diagonal form in the computational basis.
diagonalize_qwc_groupings(qwc_groupings)	Diagonalizes a list of qubit-wise commutative groupings of Pauli strings.
diagonalize_qwc_pauli_words(qwc_grouping)	Diagonalizes a list of mutually qubit-wise commutative Pauli words.
group_observables(observables[, …])	Partitions a list of observables (Pauli operations and tensor products thereof) into groupings according to a binary relation (qubit-wise commuting, fully-commuting, or anticommuting).
is_pauli_word(observable)	Checks if an observable instance is a Pauli word.
is_qwc(pauli_vec_1, pauli_vec_2)	Checks if two Pauli words in the binary vector representation are qubit-wise commutative.
observables_to_binary_matrix(observables[, …])	Converts a list of Pauli words to the binary vector representation and yields a row matrix of the binary vectors.
optimize_measurements(observables[, …])	Partitions then diagonalizes a list of Pauli words, facilitating simultaneous measurement of all observables within a partition.
pauli_to_binary(pauli_word[, n_qubits, wire_map])	Converts a Pauli word to the binary vector representation.
qwc_complement_adj_matrix(binary_observables)	Obtains the adjacency matrix for the complementary graph of the qubit-wise commutativity graph for a given set of observables in the binary representation.
qwc_rotation(pauli_operators)	Performs circuit implementation of diagonalizing unitary for a Pauli word.

Classes

PauliGroupingStrategy(observables[, …])

Class for partitioning a list of Pauli words according to some binary symmetric relation.

pennylane.grouping.graph_colouring Module

A module for heuristic algorithms for colouring Pauli graphs.

A Pauli graph is a graph where vertices represent Pauli words and edges denote if a specified symmetric binary relation (e.g., commutation) is satisfied for the corresponding Pauli words. The graph-colouring problem is to assign a colour to each vertex such that no vertices of the same colour are connected, using the fewest number of colours (lowest “chromatic number”) as possible.

Functions

largest_first(binary_observables, adj)	Performs graph-colouring using the Largest Degree First heuristic.
recursive_largest_first(binary_observables, adj)	Performs graph-colouring using the Recursive Largest Degree First heuristic.