atomiq.components.optimizers

Classes

Optimizer	An atomiq Component
BoundOptimizer	An Optimizer with restrictions on the values of the actor.
BisectionOptimizer	An optimizer that uses bisection to achieve the optimization target on the monitor

Module Contents

class atomiq.components.optimizers.Optimizer(actor_component, actor_name, monitor_component, monitor_name, *args, **kwargs)

Bases: atomiq.components.primitives.Component

An atomiq Component

Every component in atomiq inherits from this class. It provides basic functionality for automatic and recursive building and initialization of components (prepare, build, prerun). It also takes care for joining kernel invariants along the inheritance tree.

Note

The arguments parent and identifier are automatically passed to the component object by the atomiq object builder.

Parameters:

• parent -- The parent context of the component. Usually this is the experiment that uses the component

• identifier -- A unique name to identify the component.

• debug_output -- Set whether the component should show debug output. Using this switch rather than the debug kernel logger can allow the compiler to not include the debug commands in the kernel code if the output is not needed.

• actor_component (Parametrizable)

• actor_name (artiq.language.types.TStr)

• monitor_component (Measurable)

• monitor_name (artiq.language.types.TStr)

kernel_invariants

actor

monitor

actor_name

abstractmethod optimize()

class atomiq.components.optimizers.BoundOptimizer(actor_min, actor_max, *args, **kwargs)

Bases: Optimizer

An Optimizer with restrictions on the values of the actor.

Parameters:

• actor_min (artiq.language.types.TFloat) -- The minumum value the actor is allowed to take during optimization

• actor_max (artiq.language.types.TFloat) -- The maxiumum value the actor is allowed to take during optimization

kernel_invariants

actor_min

actor_max

class atomiq.components.optimizers.BisectionOptimizer(timestep=325 * us, epsilon=0.01, max_steps=20, switch_actor=False, *args, **kwargs)

Bases: BoundOptimizer

An optimizer that uses bisection to achieve the optimization target on the monitor

Parameters:

• timestep -- The time the individual steps in the algorithm take

• epsilon -- Acceptable relative difference from the target value (default 0.01, i.e. 1% of the target value)

• switch_actor (artiq.language.types.TBool) -- If True, the actor (needs to be Switchable) is switched on before the measurement of the monitor and switched off after (default False)

kernel_invariants

timestep

epsilon = 0.01

max_steps = 20

switch_actor = False

_actor_on()

_actor_off()

optimize(target)

Parameters:

target (artiq.language.types.TFloat)

Return type:

artiq.language.types.TFloat