Tabu Search

TabuSearchSolver

Classical solver using Tabu Search. Designed to integrate with the solver factory.

Signature

class TabuSearchSolver(BaseClassicalSolver):
    def solve(self) -> QUBOSolution

Description

This solver applies a Tabu Search metaheuristic to escape local minima and explore the solution space. It is suitable for solving QUBO instances classically without relying on quantum hardware. The implementation is based on TabuSearchSolver from the Ocean SDK, and returns solutions compatible with the QUBOSolution interface used across the qubo-solver package.

Fields

Field	Type	Description
use_quantum	bool	Have to be False to use a classical solver.
classical_solver_type	str	Set to "tabu_search" to use Tabu Search as the solving method.
tabu_x0	torch.Tensor	None
tabu_tenure	int	Number of iterations a move (bit flip) remains tabu.
tabu_max_no_improve	int	Maximum number of consecutive iterations without improvement before termination.
tabu_time_limit	float	Maximum execution time for Tabu Search, in seconds. If infinite, no time limit is applied.

Usage

from qubosolver import QUBOInstance
from qubosolver.solver import QuboSolver
from qubosolver.config import SolverConfig, ClassicalConfig

qubo = QUBOInstance(coefficients=[[-2.0, 1.0], [1.0, -2.0]])
config = SolverConfig(use_quantum = False, classical=ClassicalConfig(classical_solver_type="tabu_search", tabu_time_limit=300.0))

solver = QuboSolver(qubo, config)

solution = solver.solve()
print(solution)

QUBOSolution(bitstrings=tensor([[0, 1]], dtype=torch.int32), costs=tensor([-2.]), counts=tensor([1]), probabilities=tensor([1.]), solution_status=)

Notes

Recommended for classical heuristics when reproducibility and control over local search dynamics are desired.

Made with Material for MkDocs (external)