Granted Patent B2

US 12,474,959 · App. 18/345,327 · Granted Nov 18, 2025

Inventors: Ofir Ezrielev (Beer Sheva, IL); Boris Shpilyuck (Ashdod, IL); Igor Dubrovsky (Beer Sheva, IL); Nisan Haimov (Beer Sheva, IL)

Assignee: Dell Products L.P.

G06F9/50G06F9/4831G06F9/5005G06F11/3409

Abstract

Claims (73)

1 . A method for prioritizing resources in a deployment, the method comprising:

identifying an impairment of a device of a set of devices of the deployment that impacts at least one cooperative process to be performed by the deployment;

performing a temporary remediation of the device, based on the impairment, to enable performance of the cooperative process, the temporary remediation putting in place a temporary repair for the deployment that allows the performance of the cooperative process;

using a trained machine learning model to obtain prediction of a difference in an operation of the temporary remediation from operation of the device of the set of devices of the deployment while the cooperative process continues to be performed, and an uncertainty quantification for the prediction;

identifying a level of influence that the temporary repair has on the performance of the cooperative process;

establishing a prioritization for permanent repair of the device of the set of devices of the deployment based on the prediction, the uncertainty quantification, and the level of influence that the temporary repair has on the performance of the cooperative process; and

using limited resources of the deployment to permanently repair the device based on the prioritization.

2 . The method of claim 1 , further comprising:

prior to identifying the impairment:

obtaining the set of the devices of the deployment;

obtaining a digital twin for the device;

training a set of machine learning models to learn differences in operation between the set of devices of the deployment and the digital twin or to learn differences between a corrected device of the deployment and the digital twin wherein each machine learning model is trained with the corresponding device from the set of devices of the deployment and the digital twin, wherein the trained machine learning model is one of the machine learning models of the set of machine learning models; and

deploying the set of devices, the digital twin, and the set of the machine learning models.

3 . The method of claim 1 , wherein identifying the impairment of the device comprises:

noting the impairment within the device of the set of devices of the deployment that restricts normal operation of the device.

4 . The method of claim 1 , wherein performing the temporary remediation of the device comprises:

using a simulation of the device from a digital twin in place of the device of the set of devices of the deployment with the impairment to maintain operation of the cooperative process to be performed by the deployment.

5 . The method of claim 1 , wherein using the trained machine learning model comprises:

obtaining input conditions of the device of the set of devices of the deployment from a digital twin; and

ingesting the conditions into the trained machine learning model to get the difference in operation and the uncertainty quantification.

6 . The method of claim 1 , wherein identifying the level of influence comprises:

obtaining a sensitivity of the performance of the cooperative process based on the temporary repair of the device of the set of the devices.

7 . The method of claim 1 , wherein establishing the prioritization for permanent repair comprises:

ordering the devices of the set of the devices that have impairments based on a magnitude of the difference in operation and the uncertainty quantification to obtain a first ordering; and

reordering the first ordering of the devices of the set of the devices that have impairments based on the level of influence of the temporary repair to obtain a final ordering.

8 . The method of claim 7 , wherein using the limited resources comprises:

scheduling, based on the final ordering, the permanent repair of the device among the permanent repairs of the devices.

9 . A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations for prioritizing resources in a deployment, the operations comprising:

identifying an impairment of a device of a set of devices of the deployment that impacts at least one cooperative process to be performed by the deployment;

performing a temporary remediation of the device, based on the impairment, to enable performance of the cooperative process, the temporary remediation putting in place a temporary repair for the deployment that allows the performance of the cooperative process;

using a trained machine learning model to obtain prediction of a difference in an operation of the temporary remediation from operation of the device of the set of devices of the deployment while the cooperative process continues to be performed, and an uncertainty quantification for the prediction;

identifying a level of influence that the temporary repair has on the performance of the cooperative process;

establishing a prioritization for permanent repair of the device of the set of devices of the deployment based on the prediction, the uncertainty quantification, and the level of influence that the temporary repair has on the performance of the cooperative process; and

using limited resources of the deployment to permanently repair the device based on the prioritization.

10 . The non-transitory machine-readable medium of claim 9 , wherein the operations further comprise:

prior to identifying the impairment:

obtaining the set of the devices of the deployment;

obtaining a digital twin for the device;

training a set of machine learning models to learn differences in operation between the set of devices of the deployment and the digital twin, wherein each machine learning model is trained with the corresponding device from the set of devices of the deployment and the digital twin, wherein the trained machine learning model is one of the machine learning models of the set of machine learning models; and

deploying the set of devices, the digital twin, and the set of the machine learning models.

11 . The non-transitory machine-readable medium of claim 9 , wherein identifying the impairment of the device comprises:

noting the impairment within the device of the set of devices of the deployment that restricts normal operation of the device.

12 . The non-transitory machine-readable medium of claim 9 , wherein performing the temporary remediation of the device comprises:

using a simulation of the device from a digital twin in place of the device of the set of devices of the deployment with the impairment to maintain operation of the cooperative process to be performed by the deployment.

13 . The non-transitory machine-readable medium of claim 9 , wherein using the trained machine learning model comprises:

obtaining input conditions of the device of the set of devices of the deployment from a digital twin; and

ingesting the conditions into the trained machine learning model to get the difference in operation and the uncertainty quantification.

14 . The non-transitory machine-readable medium of claim 9 , wherein identifying the level of influence comprises:

obtaining a sensitivity of the performance of the cooperative process based on the temporary repair of the device of the set of the devices.

15 . A data processing system, comprising:

a processor; and

a memory coupled to the processor to store instructions, which when executed by the processor, cause the data processing system to perform operations for prioritizing resources in a deployment, the operations comprising:

identifying an impairment of a device of a set of devices of the deployment that impacts at least one cooperative process to be performed by the deployment;

performing a temporary remediation of the device, based on the impairment, to enable performance of the cooperative process, the temporary remediation putting in place a temporary repair for the deployment that allows the performance of the cooperative process;

using a trained machine learning model to obtain prediction of a difference in an operation of the temporary remediation from operation of the device of the set of devices of the deployment while the cooperative process continues to be performed, and an uncertainty quantification for the prediction;

identifying a level of influence that the temporary repair has on the performance of the cooperative process;

establishing a prioritization for permanent repair of the device of the set of devices of the deployment based on the prediction, the uncertainty quantification, and the level of influence that the temporary repair has on the performance of the cooperative process; and

using limited resources of the deployment to permanently repair the device based on the prioritization.

16 . The data processing system of claim 15 , wherein the operations further comprise:

prior to identifying the impairment:

obtaining the set of the devices of the deployment;

obtaining a digital twin for the device;

training a set of machine learning models to learn differences in operation between the set of devices of the deployment and the digital twin, wherein each machine learning model is trained with the corresponding device from the set of devices of the deployment and the digital twin, wherein the trained machine learning model is one of the machine learning models of the set of machine learning models; and

deploying the set of devices, the digital twin, and the set of the machine learning models.

17 . The data processing system of claim 15 , wherein identifying the impairment of the device comprises:

noting the impairment within the device of the set of devices of the deployment that restricts normal operation of the device.

18 . The data processing system of claim 15 , wherein performing the temporary remediation of the device comprises:

using a simulation of the device from a digital twin in place of the device of the set of devices of the deployment with the impairment to maintain operation of the cooperative process to be performed by the deployment.

19 . The data processing system of claim 15 , wherein using the trained machine learning model comprises:

obtaining input conditions of the device of the set of devices of the deployment from a digital twin; and

ingesting the conditions into the trained machine learning model to get the difference in operation and the uncertainty quantification.

20 . The data processing system of claim 15 , wherein identifying the level of influence comprises:

obtaining a sensitivity of the performance of the cooperative process based on the temporary repair of the device of the set of the devices.

Assignments (1)

Continuity (1)

References Cited (48)