Skip to main contentIBM Quantum Documentation

Build a function template for Hamiltonian simulation

Package versions

The code on this page was developed using the following requirements. We recommend using these versions or newer.

qiskit[all]~=1.2.4
qiskit-ibm-runtime~=0.32.0
qiskit-serverless~=0.18.0
qiskit-ibm-catalog~=0.2.0

By now you've seen a basic example of how to get started writing, uploading, and running a program with Qiskit Serverless. If you haven't, start with What is Qiskit Serverless? for some background information first.

The workloads you are building for your own use cases are likely not as simple as the examples previously shown. For example, you might need to carefully consider what domain-level inputs and outputs are relevant for your application, how to make sure your program can be reusable across a range of those inputs, and what kind of information you need returned during the execution of the job so you can better evaluate the progress of your workload. You might even want to incorporate a "dry run" mode so you can test the impact of a set of particular inputs before sending the resulting circuits to hardware.

A function template is an example of a realistic workload that uses a specific application domain to contextualize these aspects. It is meant to be a starting point for you to modify for your own needs so you don't have to start from scratch.

This guide demonstrates creating a function template for addressing Hamiltonian simulation problems.

Write the function template

First, write a function template for Hamiltonian simulation that uses the AQC-Tensor Qiskit addon to map the problem description to a reduced-depth circuit for execution on hardware.

Throughout, the code is saved to ./source_files/template_hamiltonian_simulation.py. This file is the function template you can upload to and run remotely with Qiskit Serverless.

Get the domain-level inputs

Start by getting the inputs for the template. This example has domain-specific inputs relevant for Hamiltonian simulation (such as the Hamiltonian and observable) and capability-specific options (such as how much you want to compress the initial layers of the Trotter circuit using AQC-Tensor, or advanced options for fine-tuning error suppression and mitigation beyond the defaults that are part of this example).

./source_files/template_hamiltonian_simulation.py
from qiskit import QuantumCircuit
from qiskit_serverless import get_arguments, save_result
 
 
# Extract parameters from arguments
#
# Do this at the top of the program so it fails early if any required arguments are missing or invalid.
 
arguments = get_arguments()
 
dry_run = arguments.get("dry_run", False)
backend_name = arguments["backend_name"]
 
aqc_evolution_time = arguments["aqc_evolution_time"]
aqc_ansatz_num_trotter_steps = arguments["aqc_ansatz_num_trotter_steps"]
aqc_target_num_trotter_steps = arguments["aqc_target_num_trotter_steps"]
 
remainder_evolution_time = arguments["remainder_evolution_time"]
remainder_num_trotter_steps = arguments["remainder_num_trotter_steps"]
 
# Stop if this fidelity is achieved
aqc_stopping_fidelity = arguments.get("aqc_stopping_fidelity", 1.0)
# Stop after this number of iterations, even if stopping fidelity is not achieved
aqc_max_iterations = arguments.get("aqc_max_iterations", 500)
 
hamiltonian = arguments["hamiltonian"]
observable = arguments["observable"]
initial_state = arguments.get("initial_state", QuantumCircuit(hamiltonian.num_qubits))
./source_files/template_hamiltonian_simulation.py (appended)
import numpy as np
import json
from mergedeep import merge
 
 
# Configure `EstimatorOptions`, to control the parameters of the hardware experiment
#
# Set default options
estimator_default_options = {
    "resilience": {
        "measure_mitigation": True,
        "zne_mitigation": True,
        "zne": {
            "amplifier": "gate_folding",
            "noise_factors": [1, 2, 3],
            "extrapolated_noise_factors": list(np.linspace(0, 3, 31)),
            "extrapolator": ["exponential", "linear", "fallback"],
        },
        "measure_noise_learning": {
            "num_randomizations": 512,
            "shots_per_randomization": 512,
        },
    },
    "twirling": {
        "enable_gates": True,
        "enable_measure": True,
        "num_randomizations": 300,
        "shots_per_randomization": 100,
        "strategy": "active",
    },
}
# Merge with user-provided options
estimator_options = merge(
    arguments.get("estimator_options", {}), estimator_default_options
)

When the function template is running, it is helpful to return information in the logs by using print statements, so that you can better evaluate the workload's progress. Following is a simple example of printing the estimator_options so there is a record of the actual Estimator options used. There are many more similar examples throughout the program to report progress during execution, including the value of the objective function during the iterative component of AQC-Tensor, and the two-qubit depth of the final instruction set architecture (ISA) circuit intended for execution on hardware.

./source_files/template_hamiltonian_simulation.py (appended)
print("estimator_options =", json.dumps(estimator_options, indent=4))

Validate the inputs

An important aspect of ensuring that the template can be reused across a range of inputs is input validation. The following code is an example of verifying that the stopping fidelity during AQC-Tensor has been specified appropriately and if not, returning an informative error message for how to fix the error.

./source_files/template_hamiltonian_simulation.py (appended)
# Perform parameter validation
 
if not 0.0 < aqc_stopping_fidelity <= 1.0:
    raise ValueError(
        f"Invalid stopping fidelity: {aqc_stopping_fidelity}.  It must be a positive float no greater than 1."
    )

Perform the Hamiltonian simulation workflow using AQC-Tensor

First, prepare a dictionary to hold all of the function template outputs. Keys will be added to this dictionary throughout the workflow, and it is returned at the end of the program.

./source_files/template_hamiltonian_simulation.py (appended)
output = {}

Step 1: Map

The AQC-Tensor optimization happens in step 1 of a Qiskit pattern. First, a target state is constructed. In this example, it is constructed from a target circuit that evolves the same Hamiltonian for the same time period as the AQC portion. Then, an ansatz is generated from an equivalent circuit but with fewer Trotter steps. In the main portion of the AQC algorithm, that ansatz is iteratively brought closer to the target state. Finally, the result is combined with the remainder of the Trotter steps needed to reach the desired evolution time.

Note the additional examples of logging incorporated in the following code.

./source_files/template_hamiltonian_simulation.py (appended)
import os
os.environ["NUMBA_CACHE_DIR"] = "/data"
 
import datetime
import quimb.tensor
from scipy.optimize import OptimizeResult, minimize
from qiskit.synthesis import SuzukiTrotter
from qiskit_addon_utils.problem_generators import generate_time_evolution_circuit
from qiskit_addon_aqc_tensor.ansatz_generation import (
    generate_ansatz_from_circuit,
    AnsatzBlock,
)
from qiskit_addon_aqc_tensor.simulation import (
    tensornetwork_from_circuit,
    compute_overlap,
)
from qiskit_addon_aqc_tensor.simulation.quimb import QuimbSimulator
from qiskit_addon_aqc_tensor.objective import OneMinusFidelity
 
print("Hamiltonian:", hamiltonian)
print("Observable:", observable)
simulator_settings = QuimbSimulator(quimb.tensor.CircuitMPS, autodiff_backend="jax")
 
# Construct the AQC target circuit
aqc_target_circuit = initial_state.copy()
if aqc_evolution_time:
    aqc_target_circuit.compose(
        generate_time_evolution_circuit(
            hamiltonian,
            synthesis=SuzukiTrotter(reps=aqc_target_num_trotter_steps),
            time=aqc_evolution_time,
        ),
        inplace=True,
    )
 
# Construct matrix-product state representation of the AQC target state
aqc_target_mps = tensornetwork_from_circuit(aqc_target_circuit, simulator_settings)
print("Target MPS maximum bond dimension:", aqc_target_mps.psi.max_bond())
output["target_bond_dimension"] = aqc_target_mps.psi.max_bond()
 
# Generate an ansatz and initial parameters from a Trotter circuit with fewer steps
aqc_good_circuit = initial_state.copy()
if aqc_evolution_time:
    aqc_good_circuit.compose(
        generate_time_evolution_circuit(
            hamiltonian,
            synthesis=SuzukiTrotter(reps=aqc_ansatz_num_trotter_steps),
            time=aqc_evolution_time,
        ),
        inplace=True,
    )
aqc_ansatz, aqc_initial_parameters = generate_ansatz_from_circuit(aqc_good_circuit)
print("Number of AQC parameters:", len(aqc_initial_parameters))
output["num_aqc_parameters"] = len(aqc_initial_parameters)
 
# Calculate the fidelity of ansatz circuit vs. the target state, before optimization
good_mps = tensornetwork_from_circuit(aqc_good_circuit, simulator_settings)
starting_fidelity = abs(compute_overlap(good_mps, aqc_target_mps)) ** 2
print("Starting fidelity of AQC portion:", starting_fidelity)
output["aqc_starting_fidelity"] = starting_fidelity
 
# Optimize the ansatz parameters by using MPS calculations
def callback(intermediate_result: OptimizeResult):
    fidelity = 1 - intermediate_result.fun
    print(f"{datetime.datetime.now()} Intermediate result: Fidelity {fidelity:.8f}")
    if intermediate_result.fun < stopping_point:
        raise StopIteration
 
 
objective = OneMinusFidelity(aqc_target_mps, aqc_ansatz, simulator_settings)
stopping_point = 1.0 - aqc_stopping_fidelity
 
result = minimize(
    objective,
    aqc_initial_parameters,
    method="L-BFGS-B",
    jac=True,
    options={"maxiter": aqc_max_iterations},
    callback=callback,
)
if result.status not in (
    0,
    1,
    99,
):  # 0 => success; 1 => max iterations reached; 99 => early termination via StopIteration
    raise RuntimeError(
        f"Optimization failed: {result.message} (status={result.status})"
    )
print(f"Done after {result.nit} iterations.")
output["num_iterations"] = result.nit
aqc_final_parameters = result.x
output["aqc_final_parameters"] = list(aqc_final_parameters)
 
# Construct an optimized circuit for initial portion of time evolution
aqc_final_circuit = aqc_ansatz.assign_parameters(aqc_final_parameters)
 
# Calculate fidelity after optimization
aqc_final_mps = tensornetwork_from_circuit(aqc_final_circuit, simulator_settings)
aqc_fidelity = abs(compute_overlap(aqc_final_mps, aqc_target_mps)) ** 2
print("Fidelity of AQC portion:", aqc_fidelity)
output["aqc_fidelity"] = aqc_fidelity
 
# Construct final circuit, with remainder of time evolution
final_circuit = aqc_final_circuit.copy()
if remainder_evolution_time:
    remainder_circuit = generate_time_evolution_circuit(
        hamiltonian,
        synthesis=SuzukiTrotter(reps=remainder_num_trotter_steps),
        time=remainder_evolution_time,
    )
    final_circuit.compose(remainder_circuit, inplace=True)

Step 2: Optimize

After the AQC portion of the workflow, the final_circuit is transpiled for the hardware as usual.

./source_files/template_hamiltonian_simulation.py (appended)
from qiskit_ibm_runtime import QiskitRuntimeService
from qiskit.transpiler.preset_passmanagers import generate_preset_pass_manager
 
service = QiskitRuntimeService()
backend = service.backend(backend_name)
 
# Transpile PUBs (circuits and observables) to match ISA
pass_manager = generate_preset_pass_manager(backend=backend, optimization_level=3)
isa_circuit = pass_manager.run(final_circuit)
isa_observable = observable.apply_layout(isa_circuit.layout)
 
isa_2qubit_depth = isa_circuit.depth(lambda x: x.operation.num_qubits == 2)
print("ISA circuit two-qubit depth:", isa_2qubit_depth)
output["twoqubit_depth"] = isa_2qubit_depth

Exit early if using dry run mode

If dry run mode has been selected, then the program is stopped before executing on hardware. This can be useful if, for example, you want first to inspect the two-qubit depth of the ISA circuit before deciding to execute on hardware.

./source_files/template_hamiltonian_simulation.py (appended)
# Exit now if dry run; don't execute on hardware
if dry_run:
    import sys
 
    print("Exiting before hardware execution since `dry_run` is True.")
    save_result(output)
    sys.exit(0)

Step 3: Execute on hardware

./source_files/template_hamiltonian_simulation.py (appended)
# ## Step 3: Execute quantum experiments on backend
from qiskit_ibm_runtime import EstimatorV2 as Estimator
 
 
estimator = Estimator(backend, options=estimator_options)
 
# Submit the underlying Estimator job. Note that this is not the
# actual function job.
job = estimator.run([(isa_circuit, isa_observable)])
print("Job ID:", job.job_id())
output["job_id"] = job.job_id()
 
# Wait until job is complete
hw_results = job.result()
hw_results_dicts = [pub_result.data.__dict__ for pub_result in hw_results]
 
# Save hardware results to serverless output dictionary
output["hw_results"] = hw_results_dicts
 
# Reorganize expectation values
hw_expvals = [pub_result_data["evs"].tolist() for pub_result_data in hw_results_dicts]
 
# Save expectation values to Qiskit Serverless
print("Hardware expectation values", hw_expvals)
output["hw_expvals"] = hw_expvals[0]

Save the output

This function template returns the relevant domain-level output for this Hamiltonian simulation workflow (expectation values) in addition to important metadata generated along the way.

./source_files/template_hamiltonian_simulation.py (appended)
save_result(output)

Deploy to IBM Quantum Platform

The previous section created a program to be run remotely. The code in this section uploads that program to Qiskit Serverless.

Use qiskit-ibm-catalog to authenticate to QiskitServerless with your API token, which you can find in your IBM Quantum™ account, and upload the program.

You can optionally use save_account() to save your credentials (see the "Authenticate to the service" step) in the Set up to use IBM Quantum Platform guide. Note that this writes your credentials to the same file as QiskitRuntimeService.save_account().

from qiskit_ibm_catalog import QiskitServerless, QiskitFunction
 
# Authenticate to the remote cluster and submit the pattern for remote execution
serverless = QiskitServerless()

This program has custom pip dependencies. Add them to a dependencies array when constructing the QiskitFunction instance:

template = QiskitFunction(
    title="template_hamiltonian_simulation",
    entrypoint="template_hamiltonian_simulation.py",
    working_dir="./source_files/",
    dependencies=[
        "qiskit-addon-utils~=0.1.0",
        "qiskit-addon-aqc-tensor[quimb-jax]~=0.1.2",
        "mergedeep==1.3.4",
    ],
)
serverless.upload(template)

Output:

QiskitFunction(template_hamiltonian_simulation)

To check if the program successfully uploaded, use serverless.list():

serverless.list()

Output:

 QiskitFunction(template_hamiltonian_simulation),

Run the function template remotely

The function template has been uploaded, so you can run it remotely with Qiskit Serverless. First, load the template by name:

template = serverless.load("template_hamiltonian_simulation")

Next, run the template with the domain-level inputs for Hamiltonian simulation. This example specifies a 50-qubit XXZ model with random couplings, and an initial state and observable.

from itertools import chain
import numpy as np
from qiskit.quantum_info import SparsePauliOp
 
L = 50
 
# Generate the edge list for this spin-chain
edges = [(i, i + 1) for i in range(L - 1)]
# Generate an edge-coloring so we can make hw-efficient circuits
edges = edges[::2] + edges[1::2]
 
# Generate random coefficients for our XXZ Hamiltonian
np.random.seed(0)
Js = np.random.rand(L - 1) + 0.5 * np.ones(L - 1)
 
hamiltonian = SparsePauliOp.from_sparse_list(
    chain.from_iterable(
        [
            [
                ("XX", (i, j), Js[i] / 2),
                ("YY", (i, j), Js[i] / 2),
                ("ZZ", (i, j), Js[i]),
            ]
            for i, j in edges
        ]
    ),
    num_qubits=L,
)
observable = SparsePauliOp.from_sparse_list(
    [("ZZ", (L // 2 - 1, L // 2), 1.0)], num_qubits=L
)
from qiskit import QuantumCircuit
 
initial_state = QuantumCircuit(L)
for i in range(L):
    if i % 2:
        initial_state.x(i)
job = template.run(
    dry_run=True,
    initial_state=initial_state,
    hamiltonian=hamiltonian,
    observable=observable,
    backend_name="ibm_fez",
    estimator_options={},
    aqc_evolution_time=0.2,
    aqc_ansatz_num_trotter_steps=1,
    aqc_target_num_trotter_steps=32,
    remainder_evolution_time=0.2,
    remainder_num_trotter_steps=4,
    aqc_max_iterations=300,
)
print(job.job_id)

Output:

853b0edb-d63f-4629-be71-398b6dcf33cb

Check the status of the job:

job.status()

Output:

'QUEUED'

After the job is running, you can fetch logs created from the print() outputs. These can provide actionable information about the progress of the Hamiltonian simulation workflow. For example, the value of the objective function during the iterative component of AQC, or the two-qubit depth of the final ISA circuit intended for execution on hardware.

print(job.logs())

Output:

No logs yet.

Block the rest of the program until a result is available. After the job is done, you can retrieve the results. These include the domain-level output of Hamiltonian simulation (expectation value) and useful metadata.

result = job.result()
 
del result[
    "aqc_final_parameters"
]  # the list is too long to conveniently display here
result

Output:

{'target_bond_dimension': 5,
 'num_aqc_parameters': 816,
 'aqc_starting_fidelity': 0.9914382555614002,
 'num_iterations': 72,
 'aqc_fidelity': 0.9998108844412502,
 'twoqubit_depth': 33}

After the job completes, the entire logging output will be available.

print(job.logs())

Output:

2024-12-17 14:50:15,580	INFO job_manager.py:531 -- Runtime env is setting up.
estimator_options = {
    "resilience": {
        "measure_mitigation": true,
        "zne_mitigation": true,
        "zne": {
            "amplifier": "gate_folding",
            "noise_factors": [
                1,
                2,
                3
            ],
            "extrapolated_noise_factors": [
                0.0,
                0.1,
                0.2,
                0.30000000000000004,
                0.4,
                0.5,
                0.6000000000000001,
                0.7000000000000001,
                0.8,
                0.9,
                1.0,
                1.1,
                1.2000000000000002,
                1.3,
                1.4000000000000001,
                1.5,
                1.6,
                1.7000000000000002,
                1.8,
                1.9000000000000001,
                2.0,
                2.1,
                2.2,
                2.3000000000000003,
                2.4000000000000004,
                2.5,
                2.6,
                2.7,
                2.8000000000000003,
                2.9000000000000004,
                3.0
            ],
            "extrapolator": [
                "exponential",
                "linear",
                "fallback"
            ]
        },
        "measure_noise_learning": {
            "num_randomizations": 512,
            "shots_per_randomization": 512
        }
    },
    "twirling": {
        "enable_gates": true,
        "enable_measure": true,
        "num_randomizations": 300,
        "shots_per_randomization": 100,
        "strategy": "active"
    }
}
Hamiltonian: SparsePauliOp(['IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXX', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYY', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZ', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'XXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'YYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'ZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXI', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYI', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZI', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IIIZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IXXIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IYYIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII', 'IZZIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII'],
              coeffs=[0.52440675+0.j, 0.52440675+0.j, 1.0488135 +0.j, 0.55138169+0.j,
 0.55138169+0.j, 1.10276338+0.j, 0.4618274 +0.j, 0.4618274 +0.j,
 0.9236548 +0.j, 0.46879361+0.j, 0.46879361+0.j, 0.93758721+0.j,
 0.73183138+0.j, 0.73183138+0.j, 1.46366276+0.j, 0.64586252+0.j,
 0.64586252+0.j, 1.29172504+0.j, 0.53402228+0.j, 0.53402228+0.j,
 1.06804456+0.j, 0.28551803+0.j, 0.28551803+0.j, 0.57103606+0.j,
 0.2601092 +0.j, 0.2601092 +0.j, 0.5202184 +0.j, 0.63907838+0.j,
 0.63907838+0.j, 1.27815675+0.j, 0.73930917+0.j, 0.73930917+0.j,
 1.47861834+0.j, 0.48073968+0.j, 0.48073968+0.j, 0.96147936+0.j,
 0.30913721+0.j, 0.30913721+0.j, 0.61827443+0.j, 0.32167664+0.j,
 0.32167664+0.j, 0.64335329+0.j, 0.51092416+0.j, 0.51092416+0.j,
 1.02184832+0.j, 0.38227781+0.j, 0.38227781+0.j, 0.76455561+0.j,
 0.47807517+0.j, 0.47807517+0.j, 0.95615033+0.j, 0.2593949 +0.j,
 0.2593949 +0.j, 0.5187898 +0.j, 0.55604786+0.j, 0.55604786+0.j,
 1.11209572+0.j, 0.72187404+0.j, 0.72187404+0.j, 1.44374808+0.j,
 0.42975395+0.j, 0.42975395+0.j, 0.8595079 +0.j, 0.5988156 +0.j,
 0.5988156 +0.j, 1.1976312 +0.j, 0.58338336+0.j, 0.58338336+0.j,
 1.16676672+0.j, 0.35519128+0.j, 0.35519128+0.j, 0.71038256+0.j,
 0.40771418+0.j, 0.40771418+0.j, 0.81542835+0.j, 0.60759468+0.j,
 0.60759468+0.j, 1.21518937+0.j, 0.52244159+0.j, 0.52244159+0.j,
 1.04488318+0.j, 0.57294706+0.j, 0.57294706+0.j, 1.14589411+0.j,
 0.6958865 +0.j, 0.6958865 +0.j, 1.391773  +0.j, 0.44172076+0.j,
 0.44172076+0.j, 0.88344152+0.j, 0.51444746+0.j, 0.51444746+0.j,
 1.02889492+0.j, 0.71279832+0.j, 0.71279832+0.j, 1.42559664+0.j,
 0.29356465+0.j, 0.29356465+0.j, 0.5871293 +0.j, 0.66630992+0.j,
 0.66630992+0.j, 1.33261985+0.j, 0.68500607+0.j, 0.68500607+0.j,
 1.37001215+0.j, 0.64957928+0.j, 0.64957928+0.j, 1.29915856+0.j,
 0.64026459+0.j, 0.64026459+0.j, 1.28052918+0.j, 0.56996051+0.j,
 0.56996051+0.j, 1.13992102+0.j, 0.72233446+0.j, 0.72233446+0.j,
 1.44466892+0.j, 0.45733097+0.j, 0.45733097+0.j, 0.91466194+0.j,
 0.63711684+0.j, 0.63711684+0.j, 1.27423369+0.j, 0.53421697+0.j,
 0.53421697+0.j, 1.06843395+0.j, 0.55881775+0.j, 0.55881775+0.j,
 1.1176355 +0.j, 0.558467  +0.j, 0.558467  +0.j, 1.116934  +0.j,
 0.59091015+0.j, 0.59091015+0.j, 1.1818203 +0.j, 0.46851598+0.j,
 0.46851598+0.j, 0.93703195+0.j, 0.28011274+0.j, 0.28011274+0.j,
 0.56022547+0.j, 0.58531893+0.j, 0.58531893+0.j, 1.17063787+0.j,
 0.31446315+0.j, 0.31446315+0.j, 0.6289263 +0.j])
Observable: SparsePauliOp(['IIIIIIIIIIIIIIIIIIIIIIIIZZIIIIIIIIIIIIIIIIIIIIIIII'],
              coeffs=[1.+0.j])
Target MPS maximum bond dimension: 5
Number of AQC parameters: 816
Starting fidelity of AQC portion: 0.9914382555614002
2024-12-17 14:52:23.400028 Intermediate result: Fidelity 0.99764093
2024-12-17 14:52:23.429669 Intermediate result: Fidelity 0.99788003
2024-12-17 14:52:23.459674 Intermediate result: Fidelity 0.99795970
2024-12-17 14:52:23.489666 Intermediate result: Fidelity 0.99799067
2024-12-17 14:52:23.518545 Intermediate result: Fidelity 0.99803401
2024-12-17 14:52:23.546952 Intermediate result: Fidelity 0.99809821
2024-12-17 14:52:23.575271 Intermediate result: Fidelity 0.99824660
2024-12-17 14:52:23.604049 Intermediate result: Fidelity 0.99845326
2024-12-17 14:52:23.632709 Intermediate result: Fidelity 0.99870497
2024-12-17 14:52:23.660527 Intermediate result: Fidelity 0.99891442
2024-12-17 14:52:23.688273 Intermediate result: Fidelity 0.99904488
2024-12-17 14:52:23.716105 Intermediate result: Fidelity 0.99914438
2024-12-17 14:52:23.744336 Intermediate result: Fidelity 0.99922827
2024-12-17 14:52:23.773399 Intermediate result: Fidelity 0.99929071
2024-12-17 14:52:23.801482 Intermediate result: Fidelity 0.99932432
2024-12-17 14:52:23.830466 Intermediate result: Fidelity 0.99936460
2024-12-17 14:52:23.860738 Intermediate result: Fidelity 0.99938891
2024-12-17 14:52:23.889958 Intermediate result: Fidelity 0.99940607
2024-12-17 14:52:23.918703 Intermediate result: Fidelity 0.99941965
2024-12-17 14:52:23.949744 Intermediate result: Fidelity 0.99944337
2024-12-17 14:52:23.980871 Intermediate result: Fidelity 0.99946875
2024-12-17 14:52:24.012124 Intermediate result: Fidelity 0.99949009
2024-12-17 14:52:24.044359 Intermediate result: Fidelity 0.99952191
2024-12-17 14:52:24.075840 Intermediate result: Fidelity 0.99953669
2024-12-17 14:52:24.106303 Intermediate result: Fidelity 0.99955242
2024-12-17 14:52:24.139329 Intermediate result: Fidelity 0.99958412
2024-12-17 14:52:24.169725 Intermediate result: Fidelity 0.99960176
2024-12-17 14:52:24.198749 Intermediate result: Fidelity 0.99961606
2024-12-17 14:52:24.227874 Intermediate result: Fidelity 0.99963811
2024-12-17 14:52:24.256818 Intermediate result: Fidelity 0.99964383
2024-12-17 14:52:24.285889 Intermediate result: Fidelity 0.99964717
2024-12-17 14:52:24.315228 Intermediate result: Fidelity 0.99966064
2024-12-17 14:52:24.345322 Intermediate result: Fidelity 0.99966517
2024-12-17 14:52:24.374921 Intermediate result: Fidelity 0.99967089
2024-12-17 14:52:24.404309 Intermediate result: Fidelity 0.99968305
2024-12-17 14:52:24.432664 Intermediate result: Fidelity 0.99968889
2024-12-17 14:52:24.461639 Intermediate result: Fidelity 0.99969997
2024-12-17 14:52:24.491244 Intermediate result: Fidelity 0.99971666
2024-12-17 14:52:24.520354 Intermediate result: Fidelity 0.99972441
2024-12-17 14:52:24.549965 Intermediate result: Fidelity 0.99973561
2024-12-17 14:52:24.583464 Intermediate result: Fidelity 0.99973811
2024-12-17 14:52:24.617537 Intermediate result: Fidelity 0.99974074
2024-12-17 14:52:24.652247 Intermediate result: Fidelity 0.99974467
2024-12-17 14:52:24.686831 Intermediate result: Fidelity 0.99974991
2024-12-17 14:52:24.725476 Intermediate result: Fidelity 0.99975230
2024-12-17 14:52:24.764637 Intermediate result: Fidelity 0.99975373
2024-12-17 14:52:24.802499 Intermediate result: Fidelity 0.99975552
2024-12-17 14:52:24.839960 Intermediate result: Fidelity 0.99975885
2024-12-17 14:52:24.877472 Intermediate result: Fidelity 0.99976469
2024-12-17 14:52:24.916233 Intermediate result: Fidelity 0.99976517
2024-12-17 14:52:24.993750 Intermediate result: Fidelity 0.99976875
2024-12-17 14:52:25.034953 Intermediate result: Fidelity 0.99976887
2024-12-17 14:52:25.076197 Intermediate result: Fidelity 0.99977244
2024-12-17 14:52:25.112340 Intermediate result: Fidelity 0.99977638
2024-12-17 14:52:25.149947 Intermediate result: Fidelity 0.99977828
2024-12-17 14:52:25.190049 Intermediate result: Fidelity 0.99978174
2024-12-17 14:52:25.310903 Intermediate result: Fidelity 0.99978222
2024-12-17 14:52:25.347512 Intermediate result: Fidelity 0.99978508
2024-12-17 14:52:25.385201 Intermediate result: Fidelity 0.99978543
2024-12-17 14:52:25.457436 Intermediate result: Fidelity 0.99978770
2024-12-17 14:52:25.497133 Intermediate result: Fidelity 0.99978818
2024-12-17 14:52:25.541179 Intermediate result: Fidelity 0.99978913
2024-12-17 14:52:25.584791 Intermediate result: Fidelity 0.99978937
2024-12-17 14:52:25.621484 Intermediate result: Fidelity 0.99979068
2024-12-17 14:52:25.655847 Intermediate result: Fidelity 0.99979211
2024-12-17 14:52:25.691710 Intermediate result: Fidelity 0.99979700
2024-12-17 14:52:25.767711 Intermediate result: Fidelity 0.99979759
2024-12-17 14:52:25.804517 Intermediate result: Fidelity 0.99979807
2024-12-17 14:52:25.839394 Intermediate result: Fidelity 0.99980236
2024-12-17 14:52:25.874438 Intermediate result: Fidelity 0.99980296
2024-12-17 14:52:25.909900 Intermediate result: Fidelity 0.99980320
2024-12-17 14:52:26.713044 Intermediate result: Fidelity 0.99980320
Done after 72 iterations.
Fidelity of AQC portion: 0.9998108844412502
ISA circuit two-qubit depth: 33
Exiting before hardware execution since `dry_run` is True.

Next steps

Recommendations

For a deeper dive into the AQC-Tensor Qiskit addon, check out the Improved Trotterized Time Evolution with Approximate Quantum Compilation tutorial or the qiskit-addon-aqc-tensor repository.

Full program source code

Here is the entire source of ./source_files/template_hamiltonian_simulation.py as one code block.

./source_files/template_hamiltonian_simulation_full.py
from qiskit import QuantumCircuit
from qiskit_serverless import get_arguments, save_result
 
 
# Extract parameters from arguments
#
# Do this at the top of the program so it fails early if any required arguments are missing or invalid.
 
arguments = get_arguments()
 
dry_run = arguments.get("dry_run", False)
backend_name = arguments["backend_name"]
 
aqc_evolution_time = arguments["aqc_evolution_time"]
aqc_ansatz_num_trotter_steps = arguments["aqc_ansatz_num_trotter_steps"]
aqc_target_num_trotter_steps = arguments["aqc_target_num_trotter_steps"]
 
remainder_evolution_time = arguments["remainder_evolution_time"]
remainder_num_trotter_steps = arguments["remainder_num_trotter_steps"]
 
# Stop if this fidelity is achieved
aqc_stopping_fidelity = arguments.get("aqc_stopping_fidelity", 1.0)
# Stop after this number of iterations, even if stopping fidelity is not achieved
aqc_max_iterations = arguments.get("aqc_max_iterations", 500)
 
hamiltonian = arguments["hamiltonian"]
observable = arguments["observable"]
initial_state = arguments.get("initial_state", QuantumCircuit(hamiltonian.num_qubits))
 
import numpy as np
import json
from mergedeep import merge
 
 
# Configure `EstimatorOptions`, to control the parameters of the hardware experiment
#
# Set default options
estimator_default_options = {
    "resilience": {
        "measure_mitigation": True,
        "zne_mitigation": True,
        "zne": {
            "amplifier": "gate_folding",
            "noise_factors": [1, 2, 3],
            "extrapolated_noise_factors": list(np.linspace(0, 3, 31)),
            "extrapolator": ["exponential", "linear", "fallback"],
        },
        "measure_noise_learning": {
            "num_randomizations": 512,
            "shots_per_randomization": 512,
        },
    },
    "twirling": {
        "enable_gates": True,
        "enable_measure": True,
        "num_randomizations": 300,
        "shots_per_randomization": 100,
        "strategy": "active",
    },
}
# Merge with user-provided options
estimator_options = merge(
    arguments.get("estimator_options", {}), estimator_default_options
)
 
print("estimator_options =", json.dumps(estimator_options, indent=4))
 
# Perform parameter validation
 
if not 0.0 < aqc_stopping_fidelity <= 1.0:
    raise ValueError(
        f"Invalid stopping fidelity: {aqc_stopping_fidelity}.  It must be a positive float no greater than 1."
    )
 
output = {}
 
import os
os.environ["NUMBA_CACHE_DIR"] = "/data"
 
import datetime
import quimb.tensor
from scipy.optimize import OptimizeResult, minimize
from qiskit.synthesis import SuzukiTrotter
from qiskit_addon_utils.problem_generators import generate_time_evolution_circuit
from qiskit_addon_aqc_tensor.ansatz_generation import (
    generate_ansatz_from_circuit,
    AnsatzBlock,
)
from qiskit_addon_aqc_tensor.simulation import (
    tensornetwork_from_circuit,
    compute_overlap,
)
from qiskit_addon_aqc_tensor.simulation.quimb import QuimbSimulator
from qiskit_addon_aqc_tensor.objective import OneMinusFidelity
 
print("Hamiltonian:", hamiltonian)
print("Observable:", observable)
simulator_settings = QuimbSimulator(quimb.tensor.CircuitMPS, autodiff_backend="jax")
 
# Construct the AQC target circuit
aqc_target_circuit = initial_state.copy()
if aqc_evolution_time:
    aqc_target_circuit.compose(
        generate_time_evolution_circuit(
            hamiltonian,
            synthesis=SuzukiTrotter(reps=aqc_target_num_trotter_steps),
            time=aqc_evolution_time,
        ),
        inplace=True,
    )
 
# Construct matrix-product state representation of the AQC target state
aqc_target_mps = tensornetwork_from_circuit(aqc_target_circuit, simulator_settings)
print("Target MPS maximum bond dimension:", aqc_target_mps.psi.max_bond())
output["target_bond_dimension"] = aqc_target_mps.psi.max_bond()
 
# Generate an ansatz and initial parameters from a Trotter circuit with fewer steps
aqc_good_circuit = initial_state.copy()
if aqc_evolution_time:
    aqc_good_circuit.compose(
        generate_time_evolution_circuit(
            hamiltonian,
            synthesis=SuzukiTrotter(reps=aqc_ansatz_num_trotter_steps),
            time=aqc_evolution_time,
        ),
        inplace=True,
    )
aqc_ansatz, aqc_initial_parameters = generate_ansatz_from_circuit(aqc_good_circuit)
print("Number of AQC parameters:", len(aqc_initial_parameters))
output["num_aqc_parameters"] = len(aqc_initial_parameters)
 
# Calculate the fidelity of ansatz circuit vs. the target state, before optimization
good_mps = tensornetwork_from_circuit(aqc_good_circuit, simulator_settings)
starting_fidelity = abs(compute_overlap(good_mps, aqc_target_mps)) ** 2
print("Starting fidelity of AQC portion:", starting_fidelity)
output["aqc_starting_fidelity"] = starting_fidelity
 
# Optimize the ansatz parameters by using MPS calculations
def callback(intermediate_result: OptimizeResult):
    fidelity = 1 - intermediate_result.fun
    print(f"{datetime.datetime.now()} Intermediate result: Fidelity {fidelity:.8f}")
    if intermediate_result.fun < stopping_point:
        raise StopIteration
 
 
objective = OneMinusFidelity(aqc_target_mps, aqc_ansatz, simulator_settings)
stopping_point = 1.0 - aqc_stopping_fidelity
 
result = minimize(
    objective,
    aqc_initial_parameters,
    method="L-BFGS-B",
    jac=True,
    options={"maxiter": aqc_max_iterations},
    callback=callback,
)
if result.status not in (
    0,
    1,
    99,
):  # 0 => success; 1 => max iterations reached; 99 => early termination via StopIteration
    raise RuntimeError(
        f"Optimization failed: {result.message} (status={result.status})"
    )
print(f"Done after {result.nit} iterations.")
output["num_iterations"] = result.nit
aqc_final_parameters = result.x
output["aqc_final_parameters"] = list(aqc_final_parameters)
 
# Construct an optimized circuit for initial portion of time evolution
aqc_final_circuit = aqc_ansatz.assign_parameters(aqc_final_parameters)
 
# Calculate fidelity after optimization
aqc_final_mps = tensornetwork_from_circuit(aqc_final_circuit, simulator_settings)
aqc_fidelity = abs(compute_overlap(aqc_final_mps, aqc_target_mps)) ** 2
print("Fidelity of AQC portion:", aqc_fidelity)
output["aqc_fidelity"] = aqc_fidelity
 
# Construct final circuit, with remainder of time evolution
final_circuit = aqc_final_circuit.copy()
if remainder_evolution_time:
    remainder_circuit = generate_time_evolution_circuit(
        hamiltonian,
        synthesis=SuzukiTrotter(reps=remainder_num_trotter_steps),
        time=remainder_evolution_time,
    )
    final_circuit.compose(remainder_circuit, inplace=True)
 
from qiskit_ibm_runtime import QiskitRuntimeService
from qiskit.transpiler.preset_passmanagers import generate_preset_pass_manager
 
service = QiskitRuntimeService()
backend = service.backend(backend_name)
 
# Transpile PUBs (circuits and observables) to match ISA
pass_manager = generate_preset_pass_manager(backend=backend, optimization_level=3)
isa_circuit = pass_manager.run(final_circuit)
isa_observable = observable.apply_layout(isa_circuit.layout)
 
isa_2qubit_depth = isa_circuit.depth(lambda x: x.operation.num_qubits == 2)
print("ISA circuit two-qubit depth:", isa_2qubit_depth)
output["twoqubit_depth"] = isa_2qubit_depth
 
# Exit now if dry run; don't execute on hardware
if dry_run:
    import sys
 
    print("Exiting before hardware execution since `dry_run` is True.")
    save_result(output)
    sys.exit(0)
 
# ## Step 3: Execute quantum experiments on backend
from qiskit_ibm_runtime import EstimatorV2 as Estimator
 
 
estimator = Estimator(backend, options=estimator_options)
 
# Submit the underlying Estimator job. Note that this is not the
# actual function job.
job = estimator.run([(isa_circuit, isa_observable)])
print("Job ID:", job.job_id())
output["job_id"] = job.job_id()
 
# Wait until job is complete
hw_results = job.result()
hw_results_dicts = [pub_result.data.__dict__ for pub_result in hw_results]
 
# Save hardware results to serverless output dictionary
output["hw_results"] = hw_results_dicts
 
# Reorganize expectation values
hw_expvals = [pub_result_data["evs"].tolist() for pub_result_data in hw_results_dicts]
 
# Save expectation values to Qiskit Serverless
print("Hardware expectation values", hw_expvals)
output["hw_expvals"] = hw_expvals[0]
 
save_result(output)
Was this page helpful?
Report a bug or request content on GitHub.