Parameter

Parameter represents a multi-dimensional array of variational parameters managed by a VQCircuit. Each element in the array is a placeholder that can be referenced by variational quantum gates or used to build arithmetic expressions via ParamExpression.

Parameters are not constructed directly by users. Instead, they are configured on a VQCircuit through set_Param and accessed through the VQCircuit.Param method.

Overview

A Parameter object stores:

• Dimension sizes -- the shape of the parameter array (e.g., [4] for 4 parameters, [2, 3] for a 2x3 grid).
• Dimension labels -- optional human-readable names for each dimension.
• Element labels -- optional names for individual parameter elements.
• Values and gradients -- bound parameter values and their computed gradients.

Methods

size

Returns the total number of parameter elements (product of all dimension sizes).

param.size() -> int

Returns: int -- the total number of parameter elements.

---

ParamExpression

ParamExpression represents an arithmetic expression built from parameter placeholders and constants. It supports addition and multiplication, enabling users to compose parameterized values for variational quantum gates beyond simple single-placeholder references.

For example, if p0 and p1 are parameter placeholders, an expression like p0 + 0.5 * p1 can be constructed and passed as the parameter of a variational gate.

Signature

ParamExpression()                  # default constructor
ParamExpression(param_expression: ParamExpression)  # copy constructor

Methods

calculate_expression_val

Evaluates the expression numerically. After calling this method, retrieve the result with get_expression_val. The parameter values must be bound (via VQCircuit.__call__) before evaluation.

expr.calculate_expression_val() -> None

---

get_expression_val

Returns the computed value of the expression. Must be called after calculate_expression_val.

expr.get_expression_val() -> float

Returns: float -- the evaluated numeric value of the expression.

---

calculate_gradient_val

Computes the gradient of the expression with respect to its constituent placeholders, given the upstream gradient dg_div_df.

expr.calculate_gradient_val(dg_div_df: float) -> None

Parameter	Type	Description
dg_div_df	float	The upstream derivative (chain rule multiplier).

---

Arithmetic Operators

ParamExpression supports the following arithmetic operations, each producing a new ParamExpression:

Operation	Signature	Description
Addition (expr + expr)	expr + other	Adds two expressions.
Addition (expr + scalar)	expr + scalar	Adds a float constant to an expression.
Reverse addition (scalar + expr)	scalar + expr	Adds an expression to a float constant.
Multiplication (expr * expr)	expr * other	Multiplies two expressions.
Multiplication (expr * scalar)	expr * scalar	Multiplies an expression by a float constant.
Reverse multiplication (scalar * expr)	scalar * expr	Multiplies a float constant by an expression.

Returns: ParamExpression -- a new expression representing the operation.

Example

from pyqpanda3.vqcircuit import VQCircuit
from pyqpanda3.core import RX
import numpy as np

vqc = VQCircuit()
vqc.set_Param([4])

# Get individual placeholders
p0 = vqc.Param([0])
p1 = vqc.Param([1])

# Build an expression: 0.5 * p0 + p1
# Note: p0 and p1 are ParamExpression objects returned by VQCircuit.Param
expr = 0.5 * p0 + p1

# Print the expression string representation
print("Expression:", str(expr))

# Use the expression in a variational gate
vqc << RX(0, expr)

# Bind parameters and evaluate the expression
vqc(np.array([1.0, 2.0, 0.0, 0.0]))
expr.calculate_expression_val()
print("Value with params [1.0, 2.0, 0.0, 0.0]:", expr.get_expression_val())

# The total parameter count is still 4 (p0, p1, p2, p3)
# but the gate parameter is a derived expression

Output:

Expression: (0.5*Param([0])+Param([1]))
Value with params [1.0, 2.0, 0.0, 0.0]: 2.5

See Also

• VQCircuit -- the variational circuit class that manages parameters
• Gradient Results -- gradient computation result classes
• DiffMethod -- differentiation method selection