garage.tf.optimizers.penalty_lbfgs_optimizer module

class PenaltyLbfgsOptimizer(max_opt_itr=20, initial_penalty=1.0, min_penalty=0.01, max_penalty=1000000.0, increase_penalty_factor=2, decrease_penalty_factor=0.5, max_penalty_itr=10, adapt_penalty=True)

Bases: object

Performs constrained optimization via penalized L-BFGS. The penalty term is adaptively adjusted to make sure that the constraint is satisfied.

constraint_val(inputs)

loss(inputs)

optimize(inputs, name=None)

update_opt(loss, target, leq_constraint, inputs, constraint_name='constraint', name=None, *args, **kwargs)

Parameters:

• loss – Symbolic expression for the loss function.
• target – A parameterized object to optimize over. It should implement methods of the garage.core.paramerized.Parameterized class.
• leq_constraint – A constraint provided as a tuple (f, epsilon), of the form f(*inputs) <= epsilon.
• inputs – A list of symbolic variables as inputs

Returns:

No return value.