Granted Patent B1

US 12,190,331 · App. 18/420,168 · Granted Jan 7, 2025

Inventors: Justine Russo (Pittsburgh, PA); Stephen Milcoff (Pittsburgh, PA)

Assignee: PITT-OHIO

G06Q30/018G06N20/00G06Q10/04

Abstract

Claims (74)

1. An apparatus for carbon emission optimization using machine-learning, wherein the apparatus comprises:

at least a processor; and

a memory communicatively connected to the at least a processor, wherein the memory contains instructions configuring the at least a processor to:

receive an integrated logistics data collection;

determine at least one projected carbon emission as a function of the integrated logistics data collection, wherein determining the at least one projected carbon emission comprises:

training a carbon emission projection model using carbon emission training data, wherein the carbon emission training data comprises a plurality of logistics datasets as input correlated to a plurality of historical carbon emissions as output, wherein training the carbon emission projection model comprises:

updating the carbon emission training data as a function of the inputs and outputs of a previous iteration of the carbon emission projection model; and

retraining the carbon emission projection model using the updated carbon emission training data; and

determining the at least one projected carbon emission as a function of the integrated logistics data collection using the trained carbon emission projection model;

generate at least one transportation plan as a function of the integrated logistics data collection and the at least one projected carbon emission;

continuously receive a current logistics datum from an external source; and

iteratively modify the at least one transportation plan based on the current logistics datum, wherein iteratively modifying the at least one transportation plan comprises:

identifying a carbon emission outlier as a function of the current logistics datum and the trained carbon emission projection model;

determining at least one carbon emission offset as a function of the carbon emission outlier; and

updating the transportation plan to incorporate the at least one carbon emission offset.

2. The apparatus of claim 1 , wherein the integrated logistic data collection comprises order data, transportation data, and environmental data.

3. The apparatus of claim 1 , wherein the integrated logistics data collection is associated with a plurality of temporal elements.

4. The apparatus of claim 1 , wherein determining the at least one projected carbon emission comprises:

transforming the integrated logistics data collection into a logistic vector space by generating a plurality of logistics vectors using a vector embedding technique; and

determining a vector distance in the logistic vector space between each logistics vector of the plurality of logistics vectors.

5. The apparatus of claim 4 , wherein identifying the carbon emission outlier comprises:

projecting the current logistics datum to the logistic vector space; and

identifying the carbon emission outlier as a function of the vector distances between the projected current logistics datum and the plurality of logistics vectors.

6. The apparatus of claim 1 , wherein iteratively modifying the at least one transportation plan comprises:

simulating the at least one transportation plan by generating a virtual scenario based on the at least one transportation plan and the current logistics datum using a simulation module.

7. The apparatus of claim 1 , wherein identifying the carbon emission outlier comprises:

detecting a carbon emission deviation as a function of the least one projected carbon emission and a plurality of historical carbon emissions using a statistical model;

comparing the detected carbon emission deviation against a deviation threshold; and

identifying the carbon emission outlier based on the comparison.

8. The apparatus of claim 1 , wherein determining the at least one carbon emission offset comprises:

classifying the carbon emission outlier into a plurality of carbon emission categories; and

selecting at least one carbon emission offset from a set of pre-defined carbon emission offsets based on the plurality of carbon emission categories.

9. The apparatus of claim 1 , wherein adjusting the transportation plan comprises:

updating the integrated logistic data collection as a function of the at least one carbon emission offset; and

retraining the carbon emission projection model using the updated integrated logistic data.

10. The apparatus of claim 1 , wherein the memory comprises instructions further configuring the at least a processor to:

transmit the at least a modified transportation plan to a user device affiliated with a user; and

display the at least a modified transportation plan at a display device of the user device.

11. A method for carbon emission optimization using machine-learning, wherein the method comprises:

receiving, using at least a processor, an integrated logistics data collection;

determining, using the at least a processor, at least one projected carbon emission as a function of the integrated logistics data collection, wherein determining the at least one projected carbon emission comprises:

training a carbon emission projection model using carbon emission training data, wherein the carbon emission training data comprises a plurality of logistics datasets as input correlated to a plurality of historical carbon emissions as output, wherein training the carbon emission projection model comprises:

updating the carbon emission training data as a function of the inputs and outputs of a previous iteration of the carbon emission projection model; and

retraining the carbon emission projection model using the updated carbon emission training data; and

determining the at least one projected carbon emission as a function of the integrated logistics data collection using the trained carbon emission projection model;

generating, using the at least a processor, at least one transportation plan as a function of the at least one projected carbon emission;

continuously receiving, using the at least a processor, a current logistics datum from an external source; and

iteratively modifying, using the at least a processor, the at least one transportation plan, wherein iteratively modifying the at least one transportation plan comprises:

identifying a carbon emission outlier as a function of the current logistics datum and the trained carbon emission projection model;

determining at least one carbon emission offset as a function of the carbon emission outlier; and

updating the transportation plan to incorporate the at least one carbon emission offset.

12. The method of claim 11 , wherein the integrated logistic data collection comprises order data, transportation data, and environmental data.

13. The method of claim 11 , wherein the integrated logistics data collection is associated with a plurality of temporal elements.

14. The method of claim 11 , wherein determining the at least one projected carbon emission comprises:

transforming the integrated logistics data collection into a logistic vector space by generating a plurality of logistics vectors using a vector embedding technique; and

determining a vector distance in the logistic vector space between each logistics vector of the plurality of logistics vectors.

15. The method of claim 14 , wherein identifying the carbon emission outlier comprises:

projecting the current logistics datum to the logistic vector space; and

identifying the carbon emission outlier as a function of the vector distances between the projected current logistics datum and the plurality of logistics vectors.

16. The method of claim 11 , wherein iteratively modifying the at least one transportation plan comprises:

simulating the at least one transportation plan by generating a virtual scenario based on the at least one transportation plan and the current logistics datum using a simulation module.

17. The method of claim 11 , wherein identifying the carbon emission outlier comprises:

detecting a carbon emission deviation as a function of the least one projected carbon emission and a plurality of historical carbon emissions using a statistical model;

comparing the detected carbon emission deviation against a deviation threshold; and

identifying the carbon emission outlier based on the comparison.

18. The method of claim 11 , wherein determining the at least one carbon emission offset comprises:

classifying the carbon emission outlier into a plurality of carbon emission categories; and

selecting at least one carbon emission offset from a set of pre-defined carbon emission offsets based on the plurality of carbon emission categories.

19. The method of claim 11 , wherein adjusting the transportation plan comprises:

updating the integrated logistic data collection as a function of the at least one carbon emission offset; and

retraining the carbon emission projection model using the updated integrated logistic data.

20. The method of claim 11 , wherein the memory comprises instructions further configuring the at least a processor to:

transmit the at least a modified transportation plan to a user device affiliated to a user; and

display the at least a modified transportation plan at a display device with the user device.

Assignments (1)

References Cited (10)

Cited By (1)

US 12,505,146