bolero.representation.DummyBehavior¶

class bolero.representation.DummyBehavior(**kwargs)[source]¶

Dummy behavior allows using environments which do not require behaviors.

Some environments (e.g. the catapult environment) do not require behavior- search to learn actual behaviors but rather only to learn parameters (velocity and angle of a shoot in case of the catapult). This behavior encapsulates the parameters learned by the optimizer and returns them via get_outputs() to the environment whenever required. It thus connects environment and optimizer directly.

__init__(**kwargs)[source]¶

can_step()¶

Returns if step() can be called again.

Returns:	can_step : bool Can we call step() again?

get_args()¶

Get parameters for this estimator.

Returns:	params : mapping of string to any Parameter names mapped to their values.

get_n_params()[source]¶

Get number of parameters.

Returns:	n_params : int Number of parameters that will be optimized.

get_outputs(outputs)[source]¶

Get outputs of the last step.

Parameters:	outputs : array-like, shape = (n_outputs,) outputs, e.g. next action, will be updated

get_params()[source]¶

Get current parameters.

Returns:	params : array-like, shape = (n_params,) Current parameters.

init(n_inputs, n_outputs)[source]¶

Initialize the behavior.

Parameters:	n_inputs : int number of inputs n_outputs : int number of outputs

reset()[source]¶

Reset behavior.

Does nothing.

set_inputs(inputs)[source]¶

Set input for the next step.

Parameters:	inputs : array-like, shape = (0,) inputs, e.g. current state of the system

set_meta_parameters(keys, meta_parameters)[source]¶

Set meta parameters (none defined for dummy behavior).

set_params(params)[source]¶

Set new parameter values.

Parameters:	params : array-like, shape = (n_params,) New parameters.

step()[source]¶

Does nothing in DummyBehavior.