MinCellValueAggregator
python
MinCellValueAggregator(current: float = 0.0, initialized: bool = False) -> NonePersistent minimum over observed values.
Used for cost-aware exploration: a cell should remember the cheapest winning route that has ever passed through it, so the cost-gap score falls as soon as a shorter route is discovered and exploration self-extinguishes there. A running average would keep the cell lit even after a cheap route exists; the minimum tracks the best-known cost directly.
Properties
value
python
value: floatMethods
record
python
record(observation: float) -> Nonestate_dict
python
state_dict() -> dict[str, Any]Values
current
python
current: float = 0.0initialized
python
initialized: bool = False