Inventory optimization based on leftover demand distribution function

• US 10,373,117 B1
• Filed: 11/15/2016
• Issued: 08/06/2019
• Est. Priority Date: 11/15/2016
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Sadighian, Ali, Awad, Hernan P., Bass, Andrew, Maggiar, Alvaro, Robinson, Stevie, Sherman, Nicholas Deming, Zhang, Chengliang
• Primary Examiner(s): Ludwig, Peter
• Assistant Examiner(s): Walker, Michael Jared
• Application Number:
  US15/352,357
• Time in Patent Office: 994 Days
• CPC Class Codes: G06Q 10/0875 Itemisation or classification of parts, supplies or services, e.g. bill of materials
• CPC Class Codes: G06Q 30/0605 Supply or demand aggregation

Abstract

Embodiments herein are directed to determining an expected demand for an item. The specialized computer system can model an impact to an expected demand for the item between a current planning horizon and a next planning horizon, including an estimated leftover demand to be fulfilled during the next planning horizon. As calculated, this leftover demand may be less than an unfulfilled demand from the current planning horizon and may be estimated based at least in part on the unfulfilled demand and on historical leftover demand between planning horizons.

Claims

1. An inventory management system comprising:

a memory configured to store computer-executable instructions; and

a processor in communication with the memory configured to execute the computer-executable instructions to at least;

receive historical leftover demand data for an item between planning horizons of the inventory management system;

receive an input associated with the item, the input describing a quantity of the item available from one or more inventory facilities associated with the inventory management system and arrival data for the item available from the one or more inventory facilities during a current planning horizon;

model an impact to an expected demand for the item between the current planning horizon and a next planning horizon, based at least in part on the input, by at least;

estimating a shortage in the quantity of the item within the current planning horizon based at least in part on a demand distribution function for the current planning horizon and on the quantity of the item,

estimating an unfulfilled demand during the current planning horizon based at least in part on the shortage, and

estimating a leftover demand to be fulfilled in the next planning horizon, the leftover demand being less than the unfulfilled demand and estimated based at least in part on the unfulfilled demand and on the historical leftover demand between the planning horizons;

compute a next quantity of the item to order for the next planning horizon based at least in part on an optimization model that utilizes a supply distribution function for the next planning horizon, the quantity of the item available from the one or more inventory facilities, and the leftover demand; and

provide, by the inventory management system, the next quantity of the item to an ordering system of the inventory management system.

2. The inventory management system of claim 1, wherein the optimization model implements a optimization formula expressed as:

3. The inventory management system of claim 1, wherein the input associated with the item is an expected date that the item arrives to the one or more inventory facilities and historical order data.

4. The inventory management system of claim 1, wherein the inventory management system further includes the ordering system and a resource system, wherein ordering system generates order plans to obtain the quantity of the item, and wherein the resource system generates resource plans to deploy resources during the next planning horizon for handling the quantity of the items.

5. A computer-implemented method, comprising:

receiving historical leftover demand data for an item over multiple planning horizons of an inventory management system;

receiving an input associated with the item, the input describing a quantity of the item available from one or more inventory facilities associated with the inventory management system and arrival data for the item available from the one or more inventory facilities during a current planning horizon;

modeling, based at least in part on the input, an impact to an expected demand for the item between the current planning horizon and a next planning horizon by at least;

estimating a shortage in the quantity of the item within the current planning horizon based at least in part on a demand distribution function for the current planning horizon and on the quantity of the item,

estimating an unfulfilled demand during the current planning horizon based at least in part on the shortage, and

estimating a leftover demand for the item based at least in part on the historical leftover demand data and the arrival data of the item, the leftover demand being less than unfulfilled demand during the current planning horizon;

estimating a quantity of the item to order for the next planning horizon based at least in part on an optimization model that utilizes a supply distribution function for the next planning horizon, the quantity of the item available from the one or more inventory facilities, and the leftover demand; and

providing the quantity of the item to order to an ordering system of the inventory management system.

6. The computer-implemented method of claim 5, further comprising:

determining an arrival distribution function based at least in part on historical arrival data; and

determining a shortage distribution function based at least in part on a current quantity of the item, the arrival distribution function, and the demand distribution function,

wherein the leftover demand is estimated based at least in part on a leftover demand distribution function, wherein the leftover demand distribution function is determined based at least in part on the shortage distribution function.

7. The computer-implemented method of claim 5, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function that indicates a likelihood of a user returning to order the item in the next planning horizon after a supply for the item is zero during the current planning horizon.

8. The computer-implemented method of claim 5, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function, wherein providing the quantity causes the ordering system to order fewer items based at least in part on output of the leftover demand distribution function during the current planning horizon.

9. The computer-implemented method of claim 5, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function, wherein a lead time of the item is variable and is an input to the leftover demand distribution function.

10. The computer-implemented method of claim 5, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function, wherein the leftover demand distribution function is based at least in part on a recourse policy.

11. The computer-implemented method of claim 5, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function, wherein the leftover demand distribution function includes national balancing.

12. The computer-implemented method of claim 5, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function, further comprising:

determining a type of the item; and

altering the leftover demand distribution function for the item based at least in part on the determination of the type of the item.

13. One or more non-transitory computer-readable storage media collectively storing computer-executable instructions that, when executed by one or more computer systems, configure the one or more computer systems to collectively perform operations comprising:

receiving historical leftover demand data for an item over multiple planning horizons of an inventory management system;

receiving an input associated with the item, the input describing a quantity of the item available from one or more inventory facilities associated with the inventory management system and arrival data for the item available from one or more inventory facilities during a current planning horizon;

modeling, based at least in part on the input, an impact to an expected demand for the item between the current planning horizon and a next planning horizon by at least;

estimating a shortage in the quantity of the item within the current planning horizon based at least in part on a demand distribution function for the current planning horizon and on the quantity of the item,

estimating an unfulfilled demand during the current planning horizon based at least in part on the shortage, and

estimating a leftover demand for the item to be fulfilled in the next planning horizon, the leftover demand being estimated based at least in part on the historical leftover demand data and the arrival data of the item, the leftover demand being less than unfulfilled demand during the current planning horizon;

computing a next quantity of the item to order for the next planning horizon based at least in part on an optimization model that utilizes a supply distribution function for the next planning horizon, the next quantity of the item available from the one or more inventory facilities, and the leftover demand; and

providing the next quantity of the item to an ordering system of the inventory management system.

14. The one or more non-transitory computer-readable storage media of claim 13, wherein the leftover demand is estimated based at least in part on a leftover demand distribution function.

15. The one or more non-transitory computer-readable storage media of claim 13, the operations further comprising:

estimating a lead time for the item based at least in part on the supply distribution function for the item,

wherein the next quantity is estimated by at least inputting the lead time and the leftover demand to the optimization model.

16. The one or more non-transitory computer-readable storage media of claim 13, wherein the optimization model alters a second quantity of items listed in a quantity order associated with the next planning horizon.

17. The one or more non-transitory computer-readable storage media of claim 13, wherein the optimization model generates additional quantities of other items associated with the next planning horizon, wherein a quantity is outputted for a particular item based at least in part on a respective leftover demand estimated for the particular item.

18. The one or more non-transitory computer-readable storage media of claim 13, wherein the leftover demand is altered based at least in part on an attrition rate.

19. The one or more non-transitory computer-readable storage media of claim 18, wherein the attrition rate varies depending upon a number of days which the item is out-of-stock.

20. The one or more non-transitory computer-readable storage media of claim 13, wherein the historical leftover demand data comprises historical supply shortage of the item and associated historical impacts to a demand for the item, and wherein the historical leftover demand data is estimated based at least in part on a likelihood of a returning demand in the next planning horizon given a current supply shortage of the item in the current planning horizon, wherein the likelihood is based at least in part on the historical supply shortage and the associated historical impacts.

21. The one or more non-transitory computer-readable storage media of claim 13, wherein the historical leftover demand data comprises a likelihood that is estimated based at least in part on a procedural or machine learning algorithm, the algorithm trained based at least in part on the historical leftover demand data.