Second Quantized Hamiltonian

You can define a chemical problem using the second quantized (fermionic) Hamiltonian instead of defining the geometry of the molecule. This approach utilizes the power of quantum chemistry software packages such as PYSCF and PSI4 to derive the most suitable Hamiltonian for the model.

Here is a short example of this method using PYSCF, which converts a molecule to a fermionic operator.

from classiq import construct_chemistry_model, execute, synthesize
from classiq.applications.chemistry import (
    ChemistryExecutionParameters,
    FermionicOperator,
    HamiltonianProblem,
    HEAParameters,
    SummedFermionicOperator,
)
from classiq.execution import OptimizerType

hamiltonian = SummedFermionicOperator(
    op_list=[
        (FermionicOperator(op_list=[("+", 0), ("-", 0)]), 0.2),
        (FermionicOperator(op_list=[("-", 1), ("-", 1)]), 0.3),
        (FermionicOperator(op_list=[("-", 2), ("-", 2)]), 0.4),
        (FermionicOperator(op_list=[("-", 3), ("-", 3)]), 0.5),
        (FermionicOperator(op_list=[("+", 0), ("+", 1), ("-", 1), ("-", 0)]), -0.1),
        (FermionicOperator(op_list=[("+", 2), ("+", 3), ("-", 2), ("-", 3)]), -0.3),
    ]
)
ham_problem = HamiltonianProblem(hamiltonian=hamiltonian, num_particles=[1, 1])

print(ham_problem.hamiltonian)

hwea_params = HEAParameters(
    num_qubits=4,
    connectivity_map=[(0, 1), (1, 2), (2, 3)],
    reps=3,
    one_qubit_gates=["x", "ry"],
    two_qubit_gates=["cx"],
)

serialized_chemistry_model = construct_chemistry_model(
    chemistry_problem=ham_problem,
    use_hartree_fock=False,
    ansatz_parameters=hwea_params,
    execution_parameters=ChemistryExecutionParameters(
        optimizer=OptimizerType.COBYLA,
        max_iteration=100,
        initial_point=None,
    ),
)

qprog = synthesize(serialized_chemistry_model)

results = execute(qprog).result()
chemistry_result = results[0].value
chemistry_result.energy

A basic script to convert MoleculeProblem to HamiltonianProblem using PYSCF is given in classiq.applications.chemistry.pyscf_hamiltonian. You can extend it to use more features from PYSCF.

The rest of the solving process is regular.