Systems and methods for simulation forecasting including dynamic realignment

Patent metadata

Field	Value
Publication number	US-12511600-B2
Application number	US-202318474925-A
Country	US
Kind code	B2
Filing date	Sep 26, 2023
Priority date	Sep 26, 2023
Publication date	Dec 30, 2025
Grant date	Dec 30, 2025

Abstract

Systems and methods for simulation forecasting in target networks including dynamic realignment are disclosed. A set of target nodes including at least one demand node realignment representative of a demand volume shift for at least one corresponding distribution node in a predetermined time period is received. When the demand volume shift is equal to or above a predetermined threshold, an estimated volume feature for the at least one corresponding distribution node is generated. When the demand volume shift is below the predetermined threshold, an actual volume feature for the at least one corresponding distribution node is generated. The generated one of the estimated volume feature or the actual volume feature is provided to a trained forecasting model to generate a demand forecast data structure based on the generated one of the estimated volume feature or the actual volume feature.

First claim

What is claimed is:

1 . A system, comprising:
a non-transitory memory; and a processor communicatively coupled to the non-transitory memory, wherein the processor is configured to read a set of instructions to:
receive a set of target nodes comprising at least one demand node realignment representative of a demand volume shift for at least one corresponding distribution node in a predetermined time period;
determine when the demand volume shift is equal to or above a predetermined volume shift threshold by:
identifying one or more realigned nodes in the set of target nodes; and
comparing a total realignment weight of the realigned nodes to the predetermined volume shift threshold;

when the demand volume shift is equal to or above the predetermined volume shift threshold, generate an estimated volume feature for the at least one corresponding distribution node by summing a projected demand volume for each demand node in a set of distribution-node specific demand nodes;
when the demand volume shift is below the predetermined volume shift threshold, generate an actual volume feature for the at least one corresponding distribution node based on historical demand period volume data for the at least one corresponding distribution node;
train a forecasting model to generate demand forecast data structures using training datasets that include at least one of aggregation data, variant-level data, holiday and event data, recall data, historical user session data, search data, purchase data, catalog data, and advertisement data;
generate, by the trained forecasting model, a demand forecast data structure based on the generated one of the estimated volume feature or the actual volume feature; and
retrain the trained forecasting model using the estimated volume feature or the actual volume feature and an associated actual demand for the predetermined time period.

2 . The system of claim 1 , wherein determining when the demand volume shift is equal to or above the predetermined volume shift threshold comprises:
determining an individual realignment weight for each realigned node in the set of target nodes; and aggregating the individual realignment weight for each realigned node to generate the total realignment weight for the set of target nodes.

3 . The system of claim 2 , wherein the individual realignment weight for each realigned node is determined based on a mean volume of each realigned node over a predetermined demand time period and a mean volume of all demand nodes over the predetermined demand time period.

4 . The system of claim 2 , wherein the individual realignment weight for each realignment node is based on a number of prior realignments of each realignment node.

5 . The system of claim 1 , wherein determining when the demand volume shift is equal to or above the predetermined volume shift threshold comprises:
generating a significant realignment response when the demand volume shift is equal to or above the predetermined volume shift threshold; and generating an insignificant realignment response when the demand volume shift is below the predetermined volume shift threshold.

6 . The system of claim 1 , wherein the estimated volume feature is generated based on historic demand volume for each demand node in the set of target demand nodes assigned to the at least one corresponding distribution node in a forecast period.

7 . The system of claim 6 , wherein the historic demand volume is obtained for historic demand periods corresponding to the forecast period in prior tracked periods.

8 . The system of claim 1 , wherein the set of target nodes are representative of a large scale distribution network.

9 . The system of claim 1 , wherein the actual volume feature is generated based on historic volume for the at least one corresponding distribution node.

10 . The system of claim 1 , wherein the processor is configured to read the set of instructions to train the forecasting model utilizing an iterative training process based, at least in part, on the generated one of the estimated volume feature or the actual volume feature.

11 . A computer-implemented method, comprising:
receiving, by a processor, a set of target nodes comprising at least one demand node realignment representative of a demand volume shift for at least one corresponding distribution node in a predetermined time period; and Determining, by the processor, when the demand volume shift is equal to or above a predetermined volume shift threshold by:
identifying one or more realigned nodes in the set of target nodes; and
comparing a total realignment weight to the predetermined volume shift threshold;
when the demand volume shift is equal to or above the predetermined volume shift threshold, generating, by the processor, an estimated volume feature for the at least one corresponding distribution node by summing a projected demand volume for each demand node in a set of distribution-node specific demand nodes; when the demand volume shift is below the predetermined volume shift threshold, generating, by the processor, an actual volume feature for the at least one corresponding distribution node based on historical demand period volume data for the at least one corresponding distribution node; training a forecasting model to generate demand forecast data structures using training datasets that include at least one of aggregation data, variant-level data, holiday and event data, recall data, historical user session data, search data, purchase data, catalog data, and advertisement data; providing, by the processor, the generated one of the estimated volume feature or the actual volume feature to the trained forecasting model; generating, by the trained forecasting model, a demand forecast data structure based on the generated one of the estimated volume feature or the actual volume feature; and retraining the trained forecasting model using the estimated volume feature or the actual volume feature and an associated actual demand for the predetermined time period.

12 . The computer-implemented method of claim 11 , wherein determining when the demand volume shift is equal to or above the predetermined volume shift threshold comprises:
determining an individual realignment weight for each realigned node in the set of target nodes; aggregating the individual realignment weight for each realigned node to generate the total realignment weight for the set of target nodes.

13 . The computer-implemented method of claim 12 , wherein the individual realignment weight for each realigned node is determined based on a mean volume of each realigned node over a predetermined demand time period and a mean volume of all demand nodes over the predetermined demand time period.

14 . The computer-implemented method of claim 12 , wherein the individual realignment weight for each realignment node is based on a number of prior realignments of each realignment node.

15 . The computer-implemented method of claim 11 , wherein determining when the demand volume shift is equal to or above the predetermined volume shift threshold comprises:
generating a significant realignment response when the demand volume shift is equal to or above the predetermined volume shift threshold; and generating an insignificant realignment response when the demand volume shift is below the predetermined volume shift threshold.

16 . The computer-implemented method of claim 11 , wherein the estimated volume feature is generated based on historic demand volume for each demand node in the set of target demand nodes assigned to the

Assignees

• Walmart Apollo Llc

Inventors

• Chowdhury Samrat
• Sanjeev Somarajan Gayathri
• Saxena Anupa
• Cao Ying
• Gupta Rajat Kumar
• Banerjee Debdut Hazra
• Zhang Jingying
• Sambasivan Lokesh Kumar

Classifications

• G06Q30/0202 Primary

  Market predictions or forecasting for commercial activities · CPC title

• G06Q10/06375

  Prediction of business process outcome or impact based on a proposed change · CPC title

• G06Q10/06315 Primary

  Needs-based resource requirements planning or analysis · CPC title