Systems and methods for use of a global registry with automated demand profiling via machine learning to optimize inventory management
An inventory machine learning process can be applied iteratively. The inventory machine learning process can be applied by a computing device. The inventory machine learning process can include determining similarity scores between demand profiles for individual parts and group centers for similarity groups of parts. The inventory machine learning process can include assigning individual parts into the similarity groups using similarity scores. The inventory machine learning process can include calculating representative demand profiles for each similarity group and determining whether the representative demand profile meets a threshold. When the threshold is met, inventory parameters can be determined for each similarity group based on the inventory machine learning process.
1 . A system for using machine learning to assign parts into similarity groups based on demand profile similarities to reduce a number of calculations and increase computational efficiency, comprising:
a data store comprising data defining a plurality of individual part demand profiles individually corresponding to a respective one of a plurality of individual parts, a plurality of similarity groups for grouping the plurality of individual parts, and a history of previous part groupings; and
at least one computing device comprising at least one processor in communication with the data store, the at least one computing device being configured to:
train an inventory machine learning process;
perform the inventory machine learning process by iteratively:
determining a respective similarity score between each respective demand profile of the plurality of individual part demand profiles and a respective group center of each of the plurality of similarity groups;
determining a count of the plurality of similarity groups in which to assign the plurality of individual parts based on the history of previous part groupings and a count of the plurality of individual parts;
performing an assignment of each of the plurality of individual parts into a respective grouping of the plurality of similarity groups based on the respective similarity score;
determining a respective profile group center as a representative demand profile for each respective similarity group; and
determining whether the calculated representative demand profile each of the plurality of similarity groups meets a predetermined group threshold;
in response to the predetermined group threshold being met, determine calculated inventory parameters for each respective similarity group based on the inventory machine learning process, wherein a count of the plurality of individual parts exceeds the count of the plurality of similarity groups; and
cause the calculated inventory parameters for each respective similarity group to be displayed on a display device.
2 . The system of claim 1 , wherein the predetermined group threshold is based on a number of iterations of the inventory machine learning process.
3 . The system of claim 1 , wherein the predetermined group threshold is based on respective similarity score between each respective demand profile of the plurality of individual part demand profiles and the respective group center of each of the plurality of similarity groups.
4 . The system of claim 1 , wherein the calculated inventory parameters for each respective similarity group are determined further based on order lead times for individual parts in each respective similarity group.
5 . The system of claim 1 , wherein the representative demand profile for each respective similarity group is calculated based on the plurality of individual part demand profiles individually corresponding to a respective one of the plurality of individual parts in the respective similarity group.
6 . The system of claim 1 , wherein the at least one computing device is further configured to determine the respective similarity score by applying a distance function between each respective profile of the plurality of individual part demand profiles and a respective group center of each of the plurality of similarity groups.
7 . The system of claim 6 , wherein the distance function comprises a Euclidean function.
8 . A non-transitory computer-readable medium for using machine learning to assign parts into similarity groups based on demand profile similarities to reduce a number of calculations and increase computational efficiency embodying a program that, when executed by at least one computing device comprising at least one processor, causes the at least one computing device to:
train an inventory machine learning process using supply chain data;
perform the inventory machine learning process by iteratively:
determining a respective similarity score between each respective demand profile of a plurality of individual part demand profiles and a respective group center of each of a plurality of similarity groups;
determining a count of the plurality of similarity groups in which to assign a plurality of individual parts based on a history of previous part groupings and a count of the plurality of individual parts;
performing an assignment of each of the plurality of individual parts into a respective grouping of the plurality of similarity groups based on the respective similarity score;
determining a respective profile group center as a representative demand profile for each respective similarity group; and
determining whether the calculated representative demand profile each of the plurality of similarity groups meets a predetermined group threshold; and
in response to the predetermined group threshold being met, determine calculated inventory parameters for each respective similarity group based on the inventory machine learning process, wherein a count of the plurality of individual parts exceeds the count of the plurality of similarity groups; and
cause the calculated inventory parameters for each respective similarity group to be displayed on a display device.
9 . The non-transitory computer-readable medium of claim 8 , wherein the representative demand profile for each respective similarity group is calculated based on a subset of the plurality of individual part demand profiles corresponding to individual parts assigned to the respective similarity group.
10 . The non-transitory computer-readable medium of claim 8 , wherein the program further causes the at least one computing device to:
in response to the calculated representative demand profile failing to meet the predetermined group threshold, calculate an updated respective group center of each of the plurality of similarity groups; and
perform another iteration.
11 . The non-transitory computer-readable medium of claim 10 , wherein the program further causes the at least one computing device to calculate the updated respective group center for each of the plurality of similarity groups by calculating a respective average of demand profiles for individual parts assigned to each of the plurality of similarity groups.
12 . The non-transitory computer-readable medium of claim 8 , wherein calculating the representative demand profile is performed in response to performing the assignment.
13 . The non-transitory computer-readable medium of claim 8 , wherein the program further causes the at least one computing device to display, on the display device, one or more of the following: one or more individual part demand profiles, one or more new individual part demand profiles, one or more calculated representative demand profiles for a respective similarity group, one or more similarity metrics, supply chain data, or new supply chain data.
14 . The non-transitory computer-readable medium of claim 8 , wherein the program further causes the at least one computing device to train the inventory machine learning process using supply chain data.
15 . The non-transitory computer-readable medium of claim 14 , wherein the program further causes the at least one computing device to train the inventory machine learning process using the supply chain data in response to the supply chain data exceeding a frequency threshold based on a policy.
16 . A method for using machine learning to assign parts into similarity groups based on demand profile similarities to reduce a number of calculations and increase computational efficiency, comprising:
performing, via at least one computing device comprising a processor, an inventory machine learning process by iteratively:
determining a respective similarity score between each respective demand profile of a plurality of individual part demand profiles and a respective group center of each of a plurality of similarity groups;
determining a count of the plurality of similarity groups in which to assign a plurality of individual parts based on a history of previous part groupings and a count of the plurality of individual parts
performing an assignment of each of the plurality of individual parts into a respective grouping of the plurality of similarity groups based on the respective similarity score;
determining a respective profile group center as a representative demand profile for each respective similarity group; and
determining whether the calculated representative demand profile each of the plurality of similarity groups meets a predetermined group threshold;
in response to the predetermined group threshold being met, determining, via the at least one computing device, calculated inventory parameters for each respective similarity group based on the calculated representative demand profile each of the plurality of similarity groups, wherein a count of the plurality of individual parts exceeds the count of the plurality of similarity groups; and
displaying, via a display device, the calculated inventory parameters for each respective similarity group.
17 . The method of claim 16 , wherein calculating the representative demand profile comprises determining an average of parameters in a subset of the plurality of individual part demand profiles corresponding to individual parts assigned to the respective similarity group.
18 . The method of claim 17 , wherein the average comprises at least one of: an arithmetic mean, a geometric mean, a median, a mode, a harmonic mean, a generalized mean, or a cubic mean.
19 . The method of claim 16 , wherein calculating the representative demand profile comprises determining an average of a plurality of line graphs corresponding to individual parts assigned to the respective similarity group.
20 . The method of claim 16 , further comprising associating, by the at least one computing device, the calculated inventory parameters for each respective similarity group with each of the plurality of individual parts assigned to its respective similarity group.