API References

idinn.sourcing_model

class idinn.sourcing_model.BaseSourcingModel(holding_cost: float, shortage_cost: float, init_inventory: float, demand_generator: UniformDemand | CustomDemand, batch_size: int, lead_time: int | None = None, regular_lead_time: int | None = None, expedited_lead_time: int | None = None, regular_order_cost: float | None = None, expedited_order_cost: float | None = None)

Bases: object

get_current_inventory() → Tensor

get_holding_cost() → float

get_init_inventory() → Tensor

get_past_demands() → Tensor

get_past_inventories() → Tensor

get_shortage_cost() → float

reset(batch_size: int | None = None) → None

class idinn.sourcing_model.DualSourcingModel(regular_lead_time: int, expedited_lead_time: int, regular_order_cost: float, expedited_order_cost: float, holding_cost: float, shortage_cost: float, init_inventory: float, demand_generator: UniformDemand | CustomDemand, batch_size: int = 1)

Bases: BaseSourcingModel

get_expedited_lead_time() → int

get_expedited_order_cost() → float

get_last_expedited_order() → Tensor

get_last_regular_order() → Tensor

get_past_expedited_orders() → Tensor

get_past_regular_orders() → Tensor

get_regular_lead_time() → int

get_regular_order_cost() → float

order(regular_q: Tensor, expedited_q: Tensor, seed: int | None = None) → None

Orders items to the inventory and update the inventory with generated demands.

Parameters:

regular_q (torch.Tensor) – The quantity of items to order from the regular supplier.
expedited_q (torch.Tensor) – The quantity of items to order from the expedited supplier.
seed (int, optional) – Random seed for reproducibility.

class idinn.sourcing_model.SingleSourcingModel(lead_time: int, holding_cost: float, shortage_cost: float, init_inventory: float, demand_generator: UniformDemand | CustomDemand, batch_size: int = 1)

Bases: BaseSourcingModel

get_lead_time() → int

get_past_orders() → Tensor

order(q: Tensor, seed: int | None = None) → None

Orders items to the inventory and update the inventory with generated demands.

Parameters:

q (torch.Tensor) – The quantity of items to order.
seed (int, optional) – Random seed for reproducibility.

idinn.single_controller

idinn.dual_controller

idinn.demand

class idinn.demand.BaseDemand

Bases: object

abstractmethod enumerate_support() → dict

abstractmethod get_max_demand() → int

abstractmethod get_min_demand() → int

abstractmethod sample(batch_size: int) → Tensor

Generate demand for one period.

Parameters:: batch_size (int) – Size of generated demands which should correspond to the batch size or the number of SKUs.

class idinn.demand.CustomDemand(demand_prob: dict[int, float])

Bases: BaseDemand

enumerate_support() → dict

get_max_demand() → int

get_min_demand() → int

sample(batch_size: int, batch_width: int = 1) → Tensor

Generate demand for one period.

Parameters:: batch_size (int) – Size of generated demands which should correspond to the batch size or the number of SKUs. If the size does not match the dimension of the elements from demand_history, demand will be upsampled or downsampled to match the size.

class idinn.demand.UniformDemand(low: int, high: int)

Bases: BaseDemand

enumerate_support() → dict

get_max_demand() → int

get_min_demand() → int

sample(batch_size: int, batch_width: int = 1) → Tensor

Generate demand for one period.

Parameters:: batch_size (int) – Size of generated demands which should correspond to the batch size or the number of SKUs.