Source code for tno.quantum.optimization.qubo.solvers._dqo._qaoa_result

"""This module contains the ``QAOAResult`` class."""

from __future__ import annotations

import itertools
from collections.abc import Mapping
from typing import TYPE_CHECKING, Any, SupportsFloat

import numpy as np
from tno.quantum.optimization.qubo.components import Freq, ResultInterface
from tno.quantum.utils.validation import check_arraylike, check_ax

if TYPE_CHECKING:
    from typing import Self

    from matplotlib.axes import Axes
    from numpy.typing import ArrayLike
    from tno.quantum.optimization.qubo.components import QUBO
    from tno.quantum.utils import BackendConfig, BitVectorLike, OptimizerConfig


[docs] class QAOAResult(ResultInterface): """Implementation of `ResultInterface` for :py:class:`QAOASolver`."""
[docs] def __init__( # noqa: PLR0913 self, best_bitvector: BitVectorLike, best_value: SupportsFloat, freq: Freq, init_beta: ArrayLike, init_gamma: ArrayLike, final_beta: ArrayLike, final_gamma: ArrayLike, expval_history: ArrayLike, training_backend: BackendConfig, evaluation_backend: BackendConfig, optimizer: OptimizerConfig, ) -> None: """Init :py:class:`QAOAResult`. Args: best_bitvector: Bitvector corresponding to the best result. best_value: Objective value of the best result. freq: Frequency object with the found energies and number of occurrences. init_beta: Initial parameters for the mixer layer. init_gamma: Initial parameters for the cost layer. final_beta: Final parameters for the mixer layer. final_gamma: Final parameters for the mixer layer. expval_history: Loss values over all optimizing iterations. training_backend: Training backend used. evaluation_backend: Evaluation backend used. optimizer: Optimizer used. """ super().__init__(best_bitvector, best_value, freq) self.init_beta = check_arraylike(init_beta, "init_beta", ndim=1) self.init_gamma = check_arraylike(init_gamma, "init_gamma", ndim=1) self.final_beta = check_arraylike(final_beta, "final_beta", ndim=1) self.final_gamma = check_arraylike(final_gamma, "final_gamma", ndim=1) self.expval_history = check_arraylike(expval_history, "expval_history", ndim=1) self.training_backend = training_backend self.evaluation_backend = evaluation_backend self.optimizer = optimizer
[docs] @classmethod def from_result( cls, qubo: QUBO, raw_result: Mapping[str, int], properties: dict[str, Any] ) -> Self: """Construct :py:class:`QAOAResult` from `raw_result` for the given `qubo`. Args: qubo: QUBO to evaluate the given bitvectors. raw_result: Mapping with bitstrings as keys and frequencies as values. properties: Dictionary containing properties used to solve QUBO. Returns: A :py:class:`QAOAResult` containing the best bitvector, best value and frequency of the best bitvector of `raw_result` based on the given `qubo`. The best bitvector has the lowest energy (value) based on the given `qubo`. When there are ties, the bitvector with the highest frequency is returned. Raises: ValueError: If `raw_result` is empty. """ freq = Freq( bitvectors=raw_result.keys(), energies=map(qubo.evaluate, raw_result.keys()), num_occurrences=raw_result.values(), ) if not freq.energies: msg = "Argument `raw_result` is empty" raise ValueError(msg) # Find the solution index with the lowest energy. Break ties by returning the # solution index with the highest number of occurrences. energies = np.array(freq.energies) num_occurrences = np.array(freq.num_occurrences) (min_indices,) = np.where(energies == energies.min()) best_idx = min_indices[np.argmax(num_occurrences[min_indices])] return cls( best_bitvector=freq.bitvectors[best_idx], best_value=freq.energies[best_idx], freq=freq, init_beta=properties["init_beta"], init_gamma=properties["init_gamma"], final_beta=properties["final_beta"], final_gamma=properties["final_gamma"], expval_history=properties["expval_history"], training_backend=properties["training_backend"], evaluation_backend=properties["evaluation_backend"], optimizer=properties["optimizer"], )
[docs] def plot_expval_history(self, ax: Axes | None = None) -> None: """Plot the history of the expectation value of the cost function. Args: ax: Optional matplotlib ``Axes`` to draw on. If ``None`` (default) create a new figure with ``Axes`` to draw on. """ ax = check_ax(ax, "ax") ax.plot(range(len(self.expval_history)), self.expval_history) ax.set_xlabel("Iteration") ax.set_ylabel("Expectation Value")
[docs] def plot_shots_histogram(self, ax: Axes | None = None) -> None: """Plot the histogram of the output of the final circuit. Args: ax: Optional matplotlib ``Axes`` to draw on. If ``None`` (default) create a new figure with ``Axes`` to draw on. """ ax = check_ax(ax, "ax") n_bits = len(self.best_bitvector) x_values = ["".join(bits) for bits in itertools.product("01", repeat=n_bits)] height = [0 for _ in x_values] for bitvector, _, n in self.freq: i = int(str(bitvector), 2) height[i] += n ax.bar(x_values, height) ax.set_xlabel("Solution") ax.set_ylabel("Number of Shots")
[docs] def plot_parameters(self, ax: Axes | None = None) -> None: """Plot the final beta and gamma parameters. Args: ax: Optional matplotlib ``Axes`` to draw on. If ``None`` (default) create a new figure with ``Axes`` to draw on. """ ax = check_ax(ax, "ax") depth = len(self.final_beta) ax.plot(range(depth), self.final_beta, label="beta") ax.plot(range(depth), self.final_gamma, label="gamma") ax.set_xlabel("Depth") ax.set_ylabel("Rotation") ax.legend()