Allocating regional inventory to reduce out-of-stock costs

• US 8,732,039 B1
• Filed: 12/29/2010
• Issued: 05/20/2014
• Est. Priority Date: 12/29/2010
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Li, Xiao Yu, Chen, Maomao
• Primary Examiner(s): Refai, Ramsey
• Application Number:
  US12/981,382
• Time in Patent Office: 1,238 Days
• Field of Search: 705/28, 705/26.81, 705/26.1, 705/26.9, 705/7.31
• US Class Current: 705/28
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 10/087 Inventory or stock management, e.g. order filling, procurement or balancing against orders
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

Methods and systems for allocating regional inventory to reduce out-of-stock costs are described. A method may include identifying a total number of units of an item to be stored in a plurality of regions and determining an order forecast for the item in each of the plurality of regions. The method may also include receiving a unit out-of-stock cost of the item in each of the plurality of regions and calculating an expected cost for each of the plurality of regions based, at least in part, on the total number of units of the item, each region'"'"'s respective order forecast, and each region'"'"'s respective unit out-of-stock cost. The method may further include allocating a portion of the total number of units of the item to each of the plurality of regions to reduce a sum of the expected costs.

Claims

1. A method, comprising:

performing, by one or more computing devices, each including at least a processor and a memory,

identifying, by at least one of the one or more computing devices, a total number of units of an item to be stored in a plurality of regions;

determining, by at least one of the one or more computing devices, an order forecast for the item in each of the plurality of regions, wherein determining the order forecast for the item includes linearly partitioning a probabilistic model of an order forecast distribution among the plurality of regions dependent upon respective parameters corresponding to the plurality of regions, wherein for each of the regions, the respective parameter includes one or more of a shape parameter or a scale parameter of the probabilistic model of the order forecast distribution;

receiving, by at least one of the one or more computing devices, a unit out-of-stock cost of the item in each of the plurality of regions, wherein for a given one of the plurality of regions, the unit out-of-stock cost of the item in the given region indicates a cost of fulfilling a unit of the item in the event that the item is out-of-stock in the given region;

calculating, by at least one of the one or more computing devices, an expected cost for each of the plurality of regions based, at least in part, on the total number of units of the item, each region'"'"'s respective order forecast, and each region'"'"'s respective unit out-of-stock cost; and

allocating, by at least one of the one or more computing devices, respective portions of the total number of units of the item to each of the plurality of regions to reduce a sum of the expected costs, such that the allocating reduces expected costs specifically due to the item being out-of-stock in the plurality of regions, wherein allocating a particular portion of the total number of units of the item to the given region comprises instructing that the particular portion of the total number of units of the item be physically stored within the given region prior to the particular portion of the total number of units being used to fulfill a customer order.

2. The method of claim 1, wherein the order forecast includes a forecast corresponding to an individual item.

3. The method of claim 1, wherein the order forecast includes a forecast corresponding to a category of items.

4. The method of claim 1, wherein identifying the total number of units of the item comprises determining a number of units of the item available to a geographic region that includes the plurality of regions.

5. The method of claim 1, wherein the probabilistic model of the order forecast distribution models the order forecast distribution as a gamma distribution.

6. The method of claim 1, wherein for a set of N regions, N being an integer greater than 1, allocating a portion of the total number of units of the item to each of the plurality of regions comprises computationally solving a set of equations given by:

C₁(1−

F₁(Rx*₁)=C₂(1−

F₂(Rx*₂)) . . . =C_N(1−

F_N(Rx*_N));

wherein R denotes the total number of units of the item to be stored, each of C₁ . . . C_N denotes the unit out-of-stock cost in a corresponding region, each of F₁ . . . F_N denotes a cumulative distribution function for a corresponding region, and each of x₁* . . . x_N* denotes a target number of units to be stored in a corresponding region.

7. The method of claim 1, wherein the unit out-of-stock cost in a given one of the plurality of regions comprises a cost of fulfilling an order for the item from another one of the plurality of regions due at least in part to the given one of the plurality of regions being unable to fulfill the order.

8. The method of claim 1, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises minimizing the sum of the expected costs.

9. The method of claim 1, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises deriving an inventory allocation factor for each of the plurality of regions that makes the expected cost for each of the plurality of regions fall within a specified range.

10. The method of claim 1, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises deriving an inventory allocation factor for each of the plurality of regions that makes the expected cost for each of the plurality of regions approximately equal to each other.

11. An article of manufacture comprising a non-transitory computer-readable storage medium having instructions stored thereon that, upon execution by a computer system, cause the computer system to perform operations comprising:

identifying a total number of units of an item to be stored in a plurality of regions;

determining an order forecast for the item in each of the plurality of regions, wherein determining the order forecast for the item includes linearly partitioning a probabilistic model of an order forecast distribution among the plurality of regions dependent upon respective parameters corresponding to the plurality of regions, wherein for each of the regions, the respective parameter includes one or more of a shape parameter or a scale parameter of the probabilistic model of the order forecast distribution;

receiving a unit out-of-stock cost of the item in each of the plurality of regions, wherein for a given one of the plurality of regions, the unit out-of-stock cost of the item in the given region indicates a cost of fulfilling a unit of the item in the event that the item is out-of-stock in the given region;

calculating an expected cost for each of the plurality of regions based, at least in part, on the total number of units of the item, each region'"'"'s respective order forecast, and each region'"'"'s respective unit out-of-stock cost; and

allocating respective portions of the total number of units of the item to each of the plurality of regions to reduce a sum of the expected costs, such that the allocating reduces expected costs specifically due to the item being out-of-stock in the plurality of regions, wherein allocating a particular portion of the total number of units of the item to the given region comprises instructing that the particular portion of the total number of units of the item be physically stored within the given region prior to the particular portion of the total number of units being used to fulfill a customer order.

12. The article of manufacture of claim 11, wherein the probabilistic model of the order forecast distribution models the order forecast distribution as a gamma distribution.

13. The article of manufacture of claim 11, wherein for a set of N regions, N being an integer greater than 1, allocating a portion of the total number of units of the item to each of the plurality of regions comprises computationally solving a set of equations given by:

C₁(1−

F₁(Rx*₁))=C₂(1−

F₂(Rx*₂)) . . . =C_N(1−

F_N(Rx*_N));

wherein R denotes the total number of units of the item to be stored, each of C₁ . . . C_N denotes the unit out-of-stock cost in a corresponding region, each of F₁ . . . F_N denotes a cumulative distribution function for a corresponding region, and each of x₁* . . . x_N* denotes a target number of units to be stored in a corresponding region.

14. The article of manufacture of claim 11, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises deriving an inventory allocation factor for each of the plurality of regions that makes the expected cost for each of the plurality of regions fall within a specified range.

15. The article of manufacture of claim 11, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises deriving an inventory allocation factor for each of the plurality of regions that makes the expected cost for each of the plurality of regions approximately equal to each other.

16. A system, comprising:

at least one processor; and

a memory coupled to the at least one processor, wherein the memory stores program instructions, and wherein the program instructions are executable by the at least one processor to perform operations including;

identifying a total number of units of an item to be stored in a plurality of regions;

determining an order forecast for the item in each of the plurality of regions, wherein determining the order forecast for the item includes linearly partitioning a probabilistic model of an order forecast distribution among the plurality of regions dependent upon respective parameters corresponding to the plurality of regions, wherein for each of the regions, the respective parameter includes one or more of a shape parameter or a scale parameter of the probabilistic model of the order forecast distribution;

receiving a unit out-of-stock cost of the item in each of the plurality of regions, wherein for a given one of the plurality of regions, the unit out-of-stock cost of the item in the given region indicates a cost of fulfilling a unit of the item in the event that the item is out-of-stock in the given region;

calculating an expected cost for each of the plurality of regions based, at least in part, on the total number of units of the item, each region'"'"'s respective order forecast, and each region'"'"'s respective unit out-of-stock cost; and

allocating respective portions of the total number of units of the item to each of the plurality of regions to reduce a sum of the expected costs, such that the allocating reduces expected costs specifically due to the item being out-of-stock in the plurality of regions, wherein allocating a particular portion of the total number of units of the item to the given region comprises instructing that the particular portion of the total number of units of the item be physically stored within the given region prior to the particular portion of the total number of units being used to fulfill a customer order.

17. The system of claim 16, wherein the probabilistic model of the order forecast distribution models the order forecast distribution as a gamma distribution.

18. The system of claim 16, wherein for a set of N regions, N being an integer greater than 1, allocating a portion of the total number of units of the item to each of the plurality of regions comprises computationally solving a set of equations given by:

C₁(1−

F₁(Rx*₁))=C₂(1−

F₂(Rx*₂)) . . . =C_N(1−

F_N(Rx*_N))

wherein R denotes the total number of units of the item to be stored, each of C₁ . . . C_N denotes the unit out-of-stock cost in a corresponding region, each of F₁ . . . F_N denotes a cumulative distribution function for a corresponding region, and each of x₁* . . . x_N* denotes a target number of units to be stored in a corresponding region.

19. The system of claim 16, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises deriving an inventory allocation factor for each of the plurality of regions that makes the expected cost for each of the plurality of regions fall within a specified range.

20. The system of claim 16, wherein allocating the portions of the total number of units to reduce the sum of the expected costs comprises deriving an inventory allocation factor for each of the plurality of regions that makes the expected cost for each of the plurality of regions approximately equal to each other.