wy_qcos.engine.qubo package
Submodules
wy_qcos.engine.qubo.qubo_precision module
- wy_qcos.engine.qubo.qubo_precision.find_matrix_gcd(matrix)
Find the greatest common divisor (GCD) of all non-zero elements.
- 参数:
matrix -- a 2D numpy array
- 返回:
the GCD of all non-zero elements in the matrix
- wy_qcos.engine.qubo.qubo_precision.scale_to_integer_matrix(matrix)
Scale matrix to interger matrix.
- 参数:
matrix -- a 2D numpy array
- 返回:
a scaled integer matrix.
- wy_qcos.engine.qubo.qubo_precision.check_matrix(matrix)
Check matrix is square array or not.
Check if the input 2D list is a square array (i.e., the number of rows and columns are equal).
- 参数:
matrix -- list(list())
- 返回:
True if valid, False otherwise error_msg: str
- 返回类型:
bool
- wy_qcos.engine.qubo.qubo_precision.check_qubo_matrix_bit_width(qubo_matrix, param_bit)
Check qubo matrix bit width.
- 参数:
qubo_matrix (np.ndarray) -- qubo matrix to be checked
param_bit (int) -- param bit width
- 返回:
success of failed (bool), error message list.
- wy_qcos.engine.qubo.qubo_precision.qubo_matrix_to_ising_matrix(qubo_matrix)
Convert QUBO matrix to ising matrix.
tip: Ising matrix is added a variable to the diagonal.
- 参数:
qubo_matrix (np.ndarray) -- qubo matrix
- 返回:
ising matrix
- 返回类型:
np.ndarray
- wy_qcos.engine.qubo.qubo_precision.ising_matrix_to_qubo_matrix(ising_matrix)
Convert ising matrix to QUBO matrix.
- 参数:
ising_matrix (np.ndarray) -- ising matrix
- 返回:
QUBO matrix
- 返回类型:
np.ndarray
- wy_qcos.engine.qubo.qubo_precision.get_spins_num(mat, max_value)
Get the number of spin variables.
Obtain the number of spin variables after the matrix is reduced in precision.
- 参数:
mat (np.ndarray) -- upper triangular matrix
max_value (int) -- max value of the matrix
- 返回:
spin variables list
- 返回类型:
list
- wy_qcos.engine.qubo.qubo_precision.precision_reduction(ising_matrix, param_bit)
Precision reduction algorithm.
- 参数:
ising_matrix (np.ndarray) -- ising matrix
param_bit (int) -- Parameters indicating matrix accuracy
- 返回:
np.ndarray: new ising matrix list: last index int: total_spins_num
- 返回类型:
tuple
- wy_qcos.engine.qubo.qubo_precision.process_qubo_solution(job_results, last_idx, qubo_matrix)
Process qubo solution.
- 参数:
job_results (dict) -- job_results returned by the driver
last_idx (list) -- last index returned by precsion reduction
qubo_matrix (np.array) -- QUBO matrix
- 返回:
new job_results
- 返回类型:
dict
wy_qcos.engine.qubo.subqubo module
- class wy_qcos.engine.qubo.subqubo.SubQUBOMultiSolution(N_I=20, N_E=1, N_S=5, qubo_matrix=None, subqubo_size=100, max_converged_num=3)
基类：
objectImplement subQUBO with multiple solution instances.
- 参数:
qubo_matrix (ndarray | None)
- set_subqubo_size(subqubo_size)
Set subqubo size.
- 参数:
subqubo_size (int) -- new_subqubo_size
- set_qubo_matrix(qubo_matrix)
Set QUBO matrix and convert to binary quadratic model.
- 参数:
qubo_matrix (np.ndarray) -- n-by-n QUBO-matrix
- get_max_converged_num()
Get max converged num.
- 返回:
converged num
- 返回类型:
max_converged_num (int)
- init_instance_pool()
Initialize the instance solutions pool include N_I solution.
- 返回:
solution pool
- 返回类型:
np.ndarray
- find_best_solution(pool)
Find the best solution.
- 返回:
best solution np.ndarray: ascending order index
- 返回类型:
- optimize_solution_pool(pool)
Optimize all solutions.
- 参数:
pool (list) -- list of solutions
- 返回:
list of optimized solutions.
- 返回类型:
list
- construct_subqubo(n_s_solutions_pool)
Construct subqubo matrix from N_S solutions pool.
- 参数:
n_s_solutions_pool (list) -- N_S solutions pool
- 返回:
subqubo matrix np.ndarray: temporary solution in QUBO-matrix np.ndarray: extracted index
- 返回类型:
np.ndarray
- merge_solution(tmp_solution, sub_solution, extracted_index)
Merge the sub solution into the tentative solution.
- 参数:
tmp_solution (QUBOSolution) -- tmp solution
sub_solution (np.ndarray) -- sub solution to be merged
extracted_index (list) -- extracted index of sub solution
- 返回:
merged solution
- 返回类型:
- create_sub_solution_pools(solution_pool)
Create N_E sub-solution pools.
Create N_E sub-solution pools from the main solution pool, each formed by randomly selecting N_S solution instances.
- 参数:
solution_pool (list) -- solutions pool.
- 返回:
A list of N_E sub solution pools.
- 返回类型:
list
- update_solution_pool(solution_pool, new_solutions)
Update the solution pool.
Update the solution pool by replacing the worst solutions with the best solution.
- 参数:
solution_pool (list) -- The current solution pool.
new_solutions (list) -- A list of new solutions to be added to the pool.
- 返回:
The best solution found in the updated pool. list: The updated solution pool.
- 返回类型:
wy_qcos.engine.qubo.tabu module
- class wy_qcos.engine.qubo.tabu.QUBOSolution(solution, energy)
基类：
objectSolution information.
- This class contains:
solution (np.ndarray): n-sized solution composed of binary variables
energy (float): energy value obtained from QUBO(QuadraticProgram) of all term
- 参数:
solution (ndarray)
energy (float)
- solution: ndarray
- energy: float
- classmethod calculate_energy(qubo, solution=None)
Calculate the energy from the QUBO-matrix & solution.
- 参数:
qubo (np.ndarray) -- n-by-n QUBO-matrix
solution (np.ndarray) -- n-sized solution composed of binary variables
- 返回:
Energy value.
- 返回类型:
float
- class wy_qcos.engine.qubo.tabu.TabuSearch(qubo, init_solution=None)
基类：
objecttabu search algorithm.
- 参数:
qubo (ndarray)
init_solution (ndarray | None)
- update_tabu(index)
Update tabu table.
- 参数:
index (int) -- index of the variable
- get_best_solution()
Obtain the optimal solution and Hamiltonian.
- 返回:
(best_solution, best_energy).
- 返回类型:
tuple
- update_impact()
Compute and update the flip cost vector for each bit.
- 返回:
- Change in objective value for
flipping each bit.
- 返回类型:
np.ndarray
- find_best_flip()
Looking for the best flip.
- 返回:
index of the best flip
- 返回类型:
int
- solve()
Solve the problem.
- 返回:
solution.
- 返回类型:
np.ndarray
- determine_tabu_tenure()
Determine an appropriate tabu tenure based on problem size.
- 返回:
Tabu tenure.
- 返回类型:
int