System and method for generating predictions for unit allocation in a materials processing network

• US 9,697,488 B1
• Filed: 03/05/2009
• Issued: 07/04/2017
• Est. Priority Date: 03/05/2009
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Shen, Jiefeng, Shaikh, Qaseem Ahmed, Weigelt, Mikael O.
• Primary Examiner(s): Mansfield, Jr., Thomas L
• Application Number:
  US12/398,602
• Time in Patent Office: 3,043 Days
• Field of Search: 705 712, 705 713, 705 722, 705 726, 705 714, 705 735, 705 737, 705 22, 705 34, 715714, 718105, 370468, 703 17
• CPC Class Codes: G06Q 10/06313 Resource planning in a project environment

Abstract

Various embodiments of a system and method for generating predictions for unit allocation in a materials processing network are described. Various embodiments may include a capacity planner for defining a model including, for each of a plurality of discrete periods of time, a corresponding representation of a materials processing network including materials handling facilities, constraints associated with ones of the materials handling facilities, and logic for determining efficient flows of materials within the materials processing network during that discrete period of time. In order to determine efficient flows of materials within the materials processing network for a particular period of time, the capacity planner may determine the discrete periods of time corresponding to representations of the model required to determine the one or more efficient flows, and evaluate the logic of the those representations without evaluating logic of representations of the model corresponding to other discrete periods of time.

Claims

1. A system, comprising:

one or more hardware processors; and

a non-transitory memory coupled to the one or more hardware processors, wherein the non-transitory memory stores program instructions executable by the one or more hardware processors to implement a capacity planner configured to;

define a model comprising, for each of a plurality of discrete periods of time;

a corresponding representation of a fulfillment network that includes one or more model components that represent multiple fulfillment centers configured to process received inventory items for subsequent fulfillment of customer orders of the inventory items,

one or more constraints that are characteristic of ones of the fulfillment centers or characteristic of the inventory items, and

logic for determining efficient flows of the inventory items within the fulfillment network during that discrete period of time, wherein the logic comprises one or more algorithms associated with respective ones of the one or more model components of the representation, and wherein the multiple fulfillment centers of the fulfillment network are maintained by a same product distributor;

in order to determine one or more particular efficient flows of the inventory items within the fulfillment network for a particular period of time, wherein at least some of the inventory items flow from one or more fulfillment centers within the fulfillment network to one or more other fulfillment centers within the fulfillment network;

determine one or more of said discrete periods of time corresponding to representations of the model required to determine said one or more particular efficient flows;

discard, based on the determined one or more discrete periods of time, one or more model components that are no longer needed for the determination of the one or more efficient flows, wherein the discarded model components correspond to other ones of the discrete periods of time, wherein based at least in part on the discarded model components the system does not need memory for storing the entire graph, and wherein memory associated with the discarded one or more model components is made available based at least in part on the discard;

determine an order in which to perform algorithms associated with one or more of the remaining model components based at least in part on interdependencies between the algorithms; and

evaluate the logic of the representations based at least in part on the determined order of the algorithms and without evaluating the discarded one or more components of the model corresponding to the other ones of the discrete periods of time; and

specify a capacity plan based at least in part on said determination of one or more particular efficient flows of the inventory items within the fulfillment network for the particular period of time, wherein said capacity plan comprises predicted allocations of inventory items among the fulfillment centers of the fulfillment network, estimated outbound shipment costs for the fulfillment network, or a recommended topology for the fulfillment network;

wherein the program instructions are further executable by the one or more hardware processors to, in accordance with the specified capacity plan, either;

cause inventory to be stocked among the fulfillment centers of the fulfillment network, or cause implementation of the recommended topology of the fulfillment network.

2. The system of claim 1, wherein said one or more constraints comprise a constraint on at least one of:

a storage capacity of one or more fulfillment centers;

or

a materials processing throughput of one or more fulfillment centers.

3. The system of claim 1, wherein the capacity planner comprises a user interface configured to initiate one or more modifications to the model.

4. The system of claim 3, wherein the capacity planner is configured to perform the determination of the one or more particular efficient flows of the inventory items based on the modified model.

5. The system of claim 3, wherein said one or more modifications comprise a modification to one or more of said representations of the fulfillment network.

6. The system of claim 5, wherein said one or more modifications comprise modifying one or more of said representations to represent one or more additional fulfillment centers.

7. The system of claim 3, wherein said one or more modifications comprise a modification or replacement of at least a portion of said logic for determining efficient flows of the inventory items within the fulfillment network.

8. The system of claim 1,

wherein at least some of said representations comprise a plurality of nodes each corresponding to a fulfillment center, and a plurality of arcs between fulfillment centers;

and

wherein each arc represents a flow of materials to be determined by the capacity planner.

9. The system of claim 1, wherein for one or more of the plurality of discrete time periods, the corresponding logic for determining efficient flows of the inventory items includes logic for determining a forecast of materials to be shipped from one or more fulfillment centers within the fulfillment network.

10. A computer-implemented method, comprising:

performing, by a capacity planner implemented via one or more computing devices comprising one or more hardware processors and non-transitory memory;

creating, by the capacity planner, a model comprising, for each of a plurality of discrete periods of time;

a corresponding representation of a fulfillment network that includes one or more model components that represent multiple fulfillment centers configured to process received inventory items for subsequent fulfillment of customer orders of the inventory items,

one or more constraints that are characteristic of ones of the fulfillment centers or characteristic of the inventory items, and

logic for determining efficient flows of the inventory items within the fulfillment network during that discrete period of time, wherein the logic comprises one or more algorithms associated with respective ones of the one or more model components of the representation, and wherein the multiple fulfillment centers of the fulfillment network are maintained by a same product distributor;

determining, by the capacity planner, one or more particular efficient flows of the inventory items within the fulfillment network for a particular period of time, wherein at least some of the inventory items flow from one or more fulfillment centers within the fulfillment network to one or more other fulfillment centers within the fulfillment network, and wherein said determining one or more particular efficient flows comprises;

determining one or more of said discrete periods of time corresponding to representations of the model required to determine said one or more particular efficient flows;

discarding, based on the determined one or more discrete periods of time, one or more model components that are no longer needed for the determination of the one or more efficient flows, wherein the discarded model components correspond to other ones of the discrete periods of time, wherein based at least in part on the discarded model components the system does not need memory for storing the entire graph, and wherein memory associated with the discarded one or more model components is made available based at least in part on the discarding;

determining an order in which to perform algorithms associated with one or more of the remaining model components based at least in part on interdependencies between the algorithms and

evaluating the logic of the representations based at least in part on the determined order of the algorithms and without evaluating the discarded one or more model components of the model corresponding to the other ones of the discrete periods of time;

wherein said creating and said determining one or more particular efficient flows are performed using one or more computer systems;

specifying, by the capacity planner, a capacity plan based at least in part on said determining one or more particular efficient flows of the inventory items within the fulfillment network for a particular period of time, wherein said capacity plan comprises predicted allocations of inventory items among the fulfillment centers of the fulfillment network, estimated outbound shipment costs for the fulfillment network, or a recommended topology for the fulfillment network; and

in accordance with the specified capacity plan;

causing inventory to be stocked among the fulfillment centers of the fulfillment network, or causing implementation of the recommended topology of the fulfillment network.

11. The computer-implemented method of claim 10, wherein said one or more constraints comprise a constraint on at least one of:

a storage capacity of one or more fulfillment centers, a materials processing throughput of one or more fulfillment centers, a cost of transporting materials, and labor for processing materials.

12. The computer-implemented method of claim 10, further comprising performing one or more modifications to the model in response to input from a user interface.

13. The computer-implemented method of claim 12, wherein the determination of the one or more particular efficient flows is based on the modified model.

14. The computer-implemented method of claim 12, wherein said one or more modifications comprise a modification to one or more of said representations of the fulfillment network.

15. The computer-implemented method of claim 14, wherein performing the one or more modification comprises modifying one or more of said representations to represent one or more additional fulfillment centers.

16. The computer-implemented method of claim 12, wherein performing said one or more modifications comprises modifying or replacing at least a portion of said logic for determining efficient flows of the inventory items within the fulfillment network.

17. The computer-implemented method of claim 10,

wherein at least some of said representations comprise a plurality of nodes each corresponding to a fulfillment center, and a plurality of arcs between fulfillment centers;

and

wherein each arc represents a flow of materials.

18. The computer-implemented method of claim 10, wherein for one or more of the plurality of discrete time periods, the corresponding logic for determining efficient flows of the inventory items includes logic for determining a forecast of materials to be shipped from one or more fulfillment centers within the fulfillment network.

19. A non-transitory computer-readable storage medium storing program instructions computer-executable by one or more hardware processors to implement a capacity planner configured to:

define a model comprising, for each of a plurality of discrete periods of time;

a corresponding representation of a fulfillment network that includes one or more model components that represent multiple fulfillment centers configured to process received inventory items for subsequent fulfillment of customer orders of the inventory items,

one or more constraints that are characteristic of ones of the fulfillment centers or characteristic of the inventory items, and

logic for determining efficient flows of the inventory items within the fulfillment network during that discrete period of time, wherein the logic comprises one or more algorithms associated with respective ones of the one or more model components of the representation, and wherein the multiple fulfillment centers of the fulfillment network are maintained by a same product distributor;

in order to determine one or more particular efficient flows of the inventory items within the fulfillment network for a particular period of time, wherein at least some of the inventory items flow from one or more fulfillment centers within the fulfillment network to one or more other fulfillment centers within the fulfillment network;

determine one or more of said discrete periods of time corresponding to representations of the model required to determine said one or more particular efficient flows;

discard, based on the determined one or more discrete periods of time, one or more model components that are no longer needed for the determination of the one or more efficient flows, wherein the discarded model components correspond to other ones of the discrete periods of time wherein based at least in part on the discarded model components the system does not need memory for storing the entire graph, and wherein memory associated with the discarded one or more model components is made available based at least in part on the discard;

determine an order in which to perform algorithms associated with one or more of the remaining model components based at least in part on interdependencies between the algorithms; and

evaluate the logic of the representations based at least in part on the determined order of the algorithms and without evaluating the discarded one or more components of the model corresponding to the other ones of the discrete periods of time; and

specify a capacity plan based at least in part on said determination of one or more particular efficient flows of the inventory items within the fulfillment network for the particular period of time, wherein said capacity plan comprises predicted allocations of inventory items among the fulfillment centers of the fulfillment network, estimated outbound shipment costs for the fulfillment network, or a recommended topology for the fulfillment network;

wherein the program instructions are further computer-executable by one or more hardware processors to, in accordance with the specified capacity plan, either;

cause inventory to be stocked among the fulfillment centers of the fulfillment network, or cause implementation of the recommended topology of the fulfillment network.

20. The non-transitory medium of claim 19, wherein said one or more constraints comprise a constraint on at least one of:

a storage capacity of one or more fulfillment centers, a materials processing throughput of one or more fulfillment centers, a cost of transporting materials, and labor for processing materials.

21. The non-transitory medium of claim 19, wherein the capacity planner comprises a user interface configured to initiate one or more modifications to the model.

22. The non-transitory medium of claim 21, wherein the capacity planner is configured to perform the determination of the one or more particular efficient flows of the inventory items based on the modified model.

23. The non-transitory medium of claim 21, wherein said one or more modifications comprise a modification to one or more of said representations of the fulfillment network.

24. The non-transitory medium of claim 23, wherein said one or more modifications comprise modifying one or more of said representations to represent one or more additional fulfillment centers.

25. The non-transitory medium of claim 21, wherein said one or more modifications comprise a modification or replacement of at least a portion of said logic for determining efficient flows of the inventory items within the fulfillment network.

26. The non-transitory medium of claim 19,

wherein at least some of said representations comprise a plurality of nodes each corresponding to a fulfillment center, and a plurality of arcs between fulfillment centers;

and

wherein each arc represents a flow of materials to be determined by the capacity planner.

27. The non-transitory medium of claim 19, wherein for one or more of the plurality of discrete time periods, the corresponding logic for determining efficient flows of the inventory items includes logic for determining a forecast of materials to be shipped from one or more fulfillment centers within the fulfillment network.