Hybrid Asynchronous Decomposition Sampler prototype framework.
beta_range
option to (Hybridized)PopulationAnnealing
and CalculateAnnealingBetaSchedule
ProgressBetaAlongSchedule
now outputs delta_beta
in addition to beta
ProcessPoolExecutor
availability optional to enable
use on systems with process limitationsEnergyWeightedResampler
PopulationAnnealing
beta
in EnergyWeightedResampler
. Use
delta_beta
instead.QPUSubproblemAutoEmbeddingSampler
(#114)ComponentDecomposer
for iterating over problem graph components (#209)greedy
-based SteepestDescentProblemSampler
/SteepestDescentSubproblemSampler
SplatComposer
was producing numpy.float64
samplesets (#235). Now it inherits dtype
from input.ICM
/IsoenergeticClusterMove
samples processor blockExplodeSamples
block that splits sampleset samples into statesHybridSampler
can now return the final state produced by the
workflow, and it can propagate sampling parameters as workflow
run-time parametersutils.flip_energy_gains
EnergyWeightedResampler
for input set re-sampling at a given
temperatureAggregatedSamples
block that aggregates or spreads input samplesConst
block that sets output state variables to constant valuesWait
block for synchronization of racing componentsPopulationAnnealing
/HybridizedPopulationAnnealing
)FixedTemperatureSampler
auto_embedding_params
argument in QPU auto-embedding samplersLoop
now terminates on EndOfStream
(alternative to Unwind
)SplatComposer
/hstack_samplesets
(act on
all samples)qpu_params
argument to sampling_params
in all QPU samplersrolling_history
in EnergyImpactDecomposer
: now
decomposing the complete problemsamples
and problem
) from State
Identity
block is now non-blocking. In racing contexts you should
now use BlockingIdentity
(or Wait
) instead.Branches
and Dup
components that generalize Parallel
Loop
, enabling run-to-energy modeKerberos
, ParallelTempering
/HybridizedParallelTempering
and SimplifiedQbsolv
RandomConstraintDecomposer
edge casesRoofDualityComposite
energy calculation