pyqpanda_alg.QmRMR.QmRMR_core¶

Classes¶

`SPSAOptimizer`	Simultaneous Perturbation Stochastic Approximation (SPSA) Optimizer.
`Feature_Selection`	Quantum-based Feature Selection using parameterized quantum circuits.

Functions¶

`plot_bar`(dic)	Plot a bar chart from a dictionary.
`plot_loss`(los)	Plot a line chart showing the loss over iterations.
`control`(qubits1, qubits2, theta)	Construct a multi-controlled RY rotation gate.

Module Contents¶

pyqpanda_alg.QmRMR.QmRMR_core.plot_bar(dic)¶

Plot a bar chart from a dictionary.

Highlights the bar with the highest value in red, while the rest are shown in blue.

Parameters¶

dicdict: A dictionary where keys are labels and values are numerical quantities to be plotted.

pyqpanda_alg.QmRMR.QmRMR_core.plot_loss(los)¶

Plot a line chart showing the loss over iterations.

Useful for visualizing the optimization process.

Parameters¶

loslist of float: A list of loss values, where each entry corresponds to one optimization step.

pyqpanda_alg.QmRMR.QmRMR_core.control(qubits1, qubits2, theta)¶

Construct a multi-controlled RY rotation gate.

Creates an RY rotation gate on the target qubit with multiple control qubits.

Parameters¶

qubits1list or Qubit: Control qubit(s). Can be a single Qubit or a list of Qubits.
qubits2Qubit: Target qubit on which the RY rotation will be applied.
thetafloat: Rotation angle for the RY gate.

Returns¶

QGate: A controlled RY gate that can be appended to a quantum circuit.

class pyqpanda_alg.QmRMR.QmRMR_core.SPSAOptimizer(objective_function, a=0.1, c=0.1, alpha=0.602, beta=0.101, max_iters=1000)¶

Simultaneous Perturbation Stochastic Approximation (SPSA) Optimizer.

This class implements the SPSA algorithm for optimizing a scalar-valued objective function that depends on multiple parameters. SPSA is useful in cases where the objective function is noisy, non-differentiable, or expensive to evaluate, such as in variational quantum algorithms.

Parameters¶

objective_functionCallable[[np.ndarray], float]: The objective function to be minimized. It must take a parameter vector as input and return a scalar loss value.
afloat, optional: Scaling factor for the learning rate (step size), default is 0.1.
cfloat, optional: Scaling factor for the perturbation, default is 0.1.
alphafloat, optional: Exponent controlling the decay rate of the learning rate, default is 0.602.
betafloat, optional: Exponent controlling the decay rate of the perturbation size, default is 0.101.
max_itersint, optional: Maximum number of iterations to perform, default is 1000.

objective_function¶

a = 0.1¶

c = 0.1¶

alpha = 0.602¶

beta = 0.101¶

max_iters = 1000¶

optimize(initial_params)¶

Run the SPSA optimization to minimize the objective function.

Parameters¶

initial_paramsnp.ndarray: Initial values for the parameters to be optimized.

Returns¶

np.ndarray: The optimized parameter vector.
list[float]: History of objective function values over iterations.

class pyqpanda_alg.QmRMR.QmRMR_core.Feature_Selection(quadratic, linear, select_num)¶

Quantum-based Feature Selection using parameterized quantum circuits.

This class builds a variational quantum circuit to model a feature selection task. The objective is to optimize the selection of a subset of features based on a given linear and quadratic objective function.

Parameters¶

quadraticnp.ndarray: Quadratic coefficient matrix used in the objective function.
linearlist[float]: Linear coefficient list used in the objective function.
select_numint: Number of features to be selected (i.e., number of bits set to 1).

Attributes¶

qb_numint: Number of qubits required, equal to the number of features.
l1list[float]: History of linear loss values during optimization.
l2list[float]: History of quadratic loss values during optimization.

>>> from pyqpanda_alg import QmRMR
>>> import numpy as np
>>> from pyqpanda3.core import *
>>> import matplotlib.pyplot as plt

>>> m = 6
>>> u = np.random.random(m)
>>> cor = np.random.random([m, m])
>>> cor = (cor.T + cor)/2
>>> ini_par = np.random.random(int(m/2)*m)*np.pi
>>> loss, choice, dic = QmRMR.Feature_Selection(cor, u, 3).get_his_res(ini_par)
>>> print(choice)
>>> plt.plot(loss)
>>> plt.show()
[0, 1, 0, 0, 1, 1]

quadratic¶

linear¶

qb_num¶

select_num¶

l1 = []¶

l2 = []¶

Circuit(qbs, para)¶

select_key(dit)¶

get_theory(para)¶

cal_loss(para)¶

get_his_res(ini_para)¶