soogo.termination module

Condition module for optimization termination criteria.

This module defines various conditions that can be used to determine when an optimization process should terminate. It includes conditions based on successful improvements, robustness of conditions, and more.

class soogo.termination.RobustCondition(termination: TerminationCondition, period=30) → None

Bases: TerminationCondition

Termination criterion that makes another termination criterion robust.

is_met() → bool

Check if the condition is met.

Return type:: bool

reset(**kwargs) → None

Reset the internal state of the condition.

Return type:: None

update(out: OptimizeResult, model: Surrogate | None = None) → None

Update the condition based on the optimization result and model.

Parameters:

out (OptimizeResult) – The optimization result containing the current state.
model (Optional[Surrogate]) – The surrogate model used in the optimization, if any. (default: None)

Return type:

None

Returns:

True if the condition is met, False otherwise.

class soogo.termination.TerminationCondition

Bases: ABC

Base class for termination conditions.

This class defines the interface for conditions that can be used to determine when an optimization process should terminate.

Parameters:

out – The optimization result containing the current state.
model – The surrogate model used in the optimization, if any.

Returns:

True if the condition is met, False otherwise.

abstractmethod is_met() → bool

Check if the condition is met.

Return type:: bool

reset(**kwargs) → None

Reset the internal state of the condition.

Return type:: None

abstractmethod update(out: OptimizeResult, model: Surrogate | None = None) → None

Update the condition based on the optimization result and model.

Parameters:

out (OptimizeResult) – The optimization result containing the current state.
model (Optional[Surrogate]) – The surrogate model used in the optimization, if any. (default: None)

Return type:

None

Returns:

True if the condition is met, False otherwise.

class soogo.termination.UnsuccessfulImprovement(threshold=0.001) → None

Bases: TerminationCondition

Condition that checks for unsuccessful improvements.

The condition is met when the relative improvement in the best objective function value is less than a specified threshold.

Parameters:: threshold – The relative improvement threshold to determine when the condition is met. (default: 0.001)

threshold: The relative improvement threshold for the condition.

value_range: The range of objective function values known so far, used to normalize the improvement check.

lowest_value: The lowest objective function value found so far in the optimization.

is_met() → bool

Check if the condition is met.

Return type:: bool

reset(keep_data_knowledge: bool = False, **kwargs) → None

Reset the internal state of the condition.

Return type:: None

update(out: OptimizeResult, model: Surrogate | None = None) → None

Update the condition based on the optimization result and model.

Parameters:

out (OptimizeResult) – The optimization result containing the current state.
model (Optional[Surrogate]) – The surrogate model used in the optimization, if any. (default: None)

Return type:

None

Returns:

True if the condition is met, False otherwise.