Encode the binary quadratic problem for submission to a given solver, using the `qp` format for data. Args: solver (:class:`dwave.cloud.solver.Solver`): The solver used. linear (dict[variable, bias]/list[variable, bias]): Linear terms of the model.
(solver: 'dwave.cloud.solver.StructuredSolver',
linear: Union[list[float], dict[int, float]],
quadratic: dict[tuple[int, int], float],
offset: float = 0,
undirected_biases: bool = False
)
| 48 | |
| 49 | |
| 50 | def encode_problem_as_qp(solver: 'dwave.cloud.solver.StructuredSolver', |
| 51 | linear: Union[list[float], dict[int, float]], |
| 52 | quadratic: dict[tuple[int, int], float], |
| 53 | offset: float = 0, |
| 54 | undirected_biases: bool = False |
| 55 | ) -> EncodedQP: |
| 56 | """Encode the binary quadratic problem for submission to a given solver, |
| 57 | using the `qp` format for data. |
| 58 | |
| 59 | Args: |
| 60 | solver (:class:`dwave.cloud.solver.Solver`): |
| 61 | The solver used. |
| 62 | |
| 63 | linear (dict[variable, bias]/list[variable, bias]): |
| 64 | Linear terms of the model. |
| 65 | |
| 66 | quadratic (dict[(variable, variable), bias]): |
| 67 | Quadratic terms of the model. |
| 68 | |
| 69 | offset (number, default=0): |
| 70 | Constant offset applied to the model. |
| 71 | |
| 72 | undirected_biases (boolean, default=False): |
| 73 | Are (quadratic) biases specified on undirected edges? |
| 74 | |
| 75 | Returns: |
| 76 | Encoded submission dictionary. |
| 77 | """ |
| 78 | # convert legacy format (list) to dict for performance |
| 79 | if isinstance(linear, abc.Sequence): |
| 80 | linear = dict(enumerate(linear)) |
| 81 | |
| 82 | active = active_qubits(linear, quadratic) |
| 83 | |
| 84 | # Encode linear terms. The coefficients of the linear terms of the objective |
| 85 | # are encoded as an array of little endian 64 bit doubles. |
| 86 | # This array is then base64 encoded into a string safe for json. |
| 87 | # The order of the terms is determined by the _encoding_qubits property |
| 88 | # specified by the server. |
| 89 | # Note: only active qubits are coded with double, inactive with NaN |
| 90 | nan = float('nan') |
| 91 | lin = [linear.get(qubit, 0 if qubit in active else nan) |
| 92 | for qubit in solver._encoding_qubits] |
| 93 | |
| 94 | lin = base64.b64encode(struct.pack('<' + ('d' * len(lin)), *lin)) |
| 95 | |
| 96 | # Encode the coefficients of the quadratic terms of the objective |
| 97 | # in the same manner as the linear terms, in the order given by the |
| 98 | # _encoding_couplers property, discarding tailing zero couplings |
| 99 | if undirected_biases: |
| 100 | # quadratic biases are given in a triangular or symmetric matrix |
| 101 | quad = [quadratic.get((q1,q2), quadratic.get((q2,q1), 0)) |
| 102 | for (q1,q2) in solver._encoding_couplers |
| 103 | if q1 in active and q2 in active] |
| 104 | else: |
| 105 | # quadratic biases are defined on directed edges, conflate with sum |
| 106 | quad = [quadratic.get((q1,q2), 0) + quadratic.get((q2,q1), 0) |
| 107 | for (q1,q2) in solver._encoding_couplers |